Carbohydrate Production in Relation to ... - Springer Link

Microb Ecol (2003) 45:237–251 DOI: 10.1007/s00248-002-2027-7 2003 Springer-Verlag New York Inc.

Carbohydrate Production in Relation to Microphytobenthic Biofilm Development: An Integrated Approach in a Tidal Mesocosm F. Orvain,1 R. Galois,1 C. Barnard,2 A. Sylvestre,1 G. Blanchard,3 P.-G. Sauriau1 1

CREMA (CNRS-IFREMER UMR 10), Centre de Recherche sur les Ecosyste`mes Marins et Aquacoles de L’Houmeau, Place du se´minaire B.P.5, 17137 L’Houmeau, France 2 De´partment de chimie-biologie, Universite´ du Que´bec a` Trois-Rivie`res, 3351 boul. des Forges, Trois-Rivie`res, Que´bec, G9A-5H7 Canada 3 LBEM (EA 3168), Universite´ de la Rochelle, Baˆtiment Marie Curie, rue Enrico Fermi, 7000 La Rochelle, France Received: 19 March 2002; Accepted: 26 September 2002; Online publication: 28 March 2003

A

B S T R A C T

Experiments were performed to evaluate short-term changes in sediment extracellular carbohydrates for a multispecific assemblage of benthic diatoms in relation to physiological status, endogenous migratory rhythms, and environmental conditions. For this purpose, a mesocosm was used, which simulated both tidal and dark: light alternating cycles under controlled conditions. Scanning electronic microscopy in combination with picture analyses indicated that natural diatom migration patterns were reproduced in the mesocosm. Two EPS fractions were operationally separated in colloidal carbohydrate measurements: alcohol-soluble EPS (termed ‘‘soluble EPS’’) and alcohol-insoluble EPS (termed ‘‘bound EPS’’). Microphytobenthic biomass followed a logistic-type curve and converged toward a maximal value termed the ‘‘biotic capacity of the local environment.’’ Both EPS fractions showed oscillations with production during photosynthetic periods and sharp decreases during night immersion periods. Productions of both EPS fractions increased with Chl a production during light periods suggesting a light dependence in relation to migratory patterns. The decreases in both EPS fractions, which occurred during night immersion periods suggest that carbohydrate hydrolysis and/or washaway affected both EPS fractions similarly in benthic environments. Our results confirm the theory according to which the two distinct fractions are under different metabolic controls. No change in soluble EPS release was obtained during the transition from logarithmic to stationary phase. On the other hand, a metabolism modification of microalgae, probably related to ammonium depletion, occurred when cells entered the stationary phase, since there was a high enhancement in bound EPS production. Mesocosm results can serve as a system of reference useful to characterize biofilm development in field investigations and to revisit the effective implication of each EPS fraction in sediment stability. Correspondence to: P.-G. Sauriau; E-mail: [email protected]

238

Introduction Epipelic diatoms are the dominant component of microphytobenthic assemblages in the Marennes-Ole´ron intertidal mudflats [9], as in many other intertidal cohesive sediments [1, 6, 10, 38, 40]. It has been demonstrated that they release exopolysaccharide exudates to facilitate movement [11, 13, 31] and that this process is involved in the rhythmic migration of epipelic diatoms in response to both tidal and diurnal cycles [6, 19, 23, 24, 27, 28, 30]. Because of their high abundance, secreted carbohydrates play an important role in intertidal sediments since they represent an important labile source of food for bacteria [15, 41] and invertebrates [10, 11] and since they have been demonstrated to be involved in sediment transport by affecting mud stability [17, 22, 24, 36, 40, 44]. Positive correlations between colloidal carbohydrates and chlorophyll a concentrations have often been reported from in situ studies [14, 38–40, 43]. However, sediment colloidal carbohydrates measured by the phenol sulfuric acid assay are extractable in water and were shown to vary in composition and size distribution [7, 8]. Commonly, carbohydrates are operationally divided in two major types: low molecular weight (LMW) sugar unit molecules (alcohol-soluble) and high molecular weight (HMW) polymeric molecules (alcohol-insoluble) [6–8, 11, 28, 31– 35, 37, 39, 44]. With reference to de Brouwer and Stal [7], we used the term of ‘‘soluble EPS’’ for the colloidal alcohol-soluble carbohydrates and ‘‘bound EPS’’ for the colloidal alcohol-insoluble carbohydrates. Differences in composition between soluble and bound EPS fractions suggest that their production is under different metabolic controls [7, 32]. Short-term dynamics in microphytobenthos biomass are now rather well understood [5, 6, 18, 27, 30, 44] but the question of the relationship between diatom biomass and both EPS fractions is still partially unresolved. The production of both EPS fractions has been found to depend on migratory patterns [28, 30–32], photosynthetic processes [6, 7, 35], nutrient depletion [33], and physiological status of microphytobenthic biofilm [7, 8, 29, 32–34, 36, 40, 44], all of these effects being under the control of environmental conditions. The utilisation of carbohydrates by bacteria [15, 41, 44] and diatoms, which are capable of heterotrophy [7, 8, 35], still reinforces the complexity in the general interpretation of exopolymer production dynamics.

F. Orvain et al.

Experiments were performed to evaluate short-term changes in extracellular carbohydrates in relation to microphytobenthic biomass development, migratory rhythm patterns and nutrient availability. For this purpose, a mesocosm was used, which simulated both tidal and dark:light alternating cycles under controlled conditions. Mesocosm studies provide appropriate tool for studying microphytobenthic behavior of multispecific assemblage as they remove many of the confounding variables seen in the field by providing a realistic benthic habitat where migratory patterns can be induced. Estimates of sediment surface covered by diatoms were assessed using scanning electronic microscopy (SEM), to check whether microphytobenthic migratory rhythms induced by experimental tidal and diurnal cycles were similar to those described under natural fluctuations [16, 24, 26, 28].

Materials and Methods Experimental procedure Silty mud was collected in the Aiguillon Cove (SW of France). One part of the sediment was sieved (1 mm) to remove macrofauna and homogenized. The other part of the sediment was evenly spread in a tray and covered with a 63-lm nylon net. After 24–36 h (i.e., the time of natural low tide) under artificial light, epipelic diatoms had migrated through the net where they accumulated. They were then collected in prefiltered seawater (Whatman GF/F) and left to settle for 1 h. Subsequently, the overlying seawater was discarded, after which suspension was mixed with sieved sediment which was placed in two tanks. Both sediment tanks were placed in the mesocosm (Fig. 1). This procedure enabled us to repeat sediment cultures growing in parallel and to duplicate the complete experimental setup. In the mesocosm, diurnal and tidal cycles were artificially simulated without strong currents to prevent erosion of surficial sediment and associated microalgae during simulated flooding and receding events. Only one 6-h daytime air exposure period was reproduced in the mesocosm between 8 AM and 2 PM over 10 successive days. Outside of these times, sediments were covered with seawater in the dark. At the sediment surface, the photosynthetically active radiation (PAR) during the simulated diurnal low tide was 140 lmol m)2 s)1, which is a nonsaturating light intensity for photosynthesis [4]. During the experiment, replicate samples (n = 3) were taken daily in each of the two tanks at 8.05 AM (5 min after the beginning of the daytime air exposure period) and at 1.55 PM (5 min before tidal immersion). Tanks were divided into 96 small compartments, which were separated by physical barriers (PVC walls). The total independence between compartments enabled us to consider samples as true replicates. Three cores were taken from two randomly chosen compartments with syringes (3 cm in

Development of a Microphytobenthic Biofilm

239

Fig. 1. Schematic representation of the mesocosm used for sediment microphytobenthos cultures. This recirculating system was equipped with pumps and artificial lights regulated by automatic clocks to simulate an experimental diurnal low-tide period once a day (from 8:00 AM to 2:00 PM).

diameter) to measure chlorophyll a and carbohydrate contents, separately. Pigment and carbohydrate contents were measured on the uppermost centimeter, which represents a greater section than the maximal layer of 2 mm where diatoms have been reported to be photosynthetically active in the field [20]. Pigment contents were also measured on the 1–2 cm layer, which provided a dark control. After sampling, sediments were immediately frozen, kept in the dark at )20C, and freeze-dried until further analyses of pigments and carbohydrates. The sediment was sampled for moisture contents (n = 2). Cores were sliced at depth intervals of 0–1, 1–2 and 2–3 cm. Sediment moisture content (g water/g dry sediment ·100) was calculated by weight difference between fresh and dried sediment (72 h, 60C). Cores (n = 2) were also taken from two different compartments on the first and last day of the experiment to measure urea, phosphate, nitrate, nitrite and ammonium concentrations using the same sampling procedure as for pigments and carbohydrates. Interstitial waters were extracted from sediment samples after centrifugation (10 min, 3000 g) and kept frozen at )20C until analyses were carried out using a Skalar autoanalyzer [2]. During the course of the experiment, a fluorimeter connected to a multiparametric probe (MARTEC) was present in the recuperation tank (see Fig. 1) to measure fluorescence due to diatom resuspension when the artificial tides rose. A thermic probe (Prosensor, Stowaway, accuracy ±0.1C) was placed into the uppermost centimeter of the sediment to record temperature every 15 min. Sediment temperature fluctuated within a range

from 15.3 ± 0.3 to 16.6 ± 0.1C for immersed and air exposure periods, respectively.

Pigment and Carbohydrate Analyses Microphytobenthos biomass was assessed by measuring the chlorophyll a content. Chloropigments were extracted from 100 mg freeze-dried sediment subsamples with 90% methanol buffered with 10% ammonium acetate for 1 h at 4C in the dark. During extraction, the samples were sonicated. After centrifugation (5 min, 2000 g), fluorescence of the supernatant was measured using a Turner Fluorometer before and after acidification (10 lL HCl 0.3 M for 1 mL MeOH). Total chlorophyll a and pheopigments were calculated according to Lorenzen’s equations [21]. Colloidal carbohydrates were extracted from 500 mg of freeze-dried sediment with 5 mL distilled water for 1 h at 30C, under continuous agitation. Samples were then centrifuged (10 min, 2000 g) and 1 mL of the resulting supernatant was used for analysis of colloidal carbohydrates. Another 2 mL aliquot of supernatant was evaporated at 60C in a dry bath, under a flow of nitrogen to separate soluble and bound EPS fractions. The residue was dissolved in 1 mL distilled water and mixed with 4 mL absolute ethanol. This mixture was left overnight at 4C, then centrifuged (15 min, 4000 g) to concentrate the precipitated polymeric fraction. The whole ethanolic supernatant was separated from the pellet and evaporated at 60C under a

240 flow of nitrogen then, the resulting pellet was dissolved in 1 mL distilled water to measure soluble EPS. The ethanol insoluble material (i.e., the pellet resulting from the precipitation) was dissolved in 1 mL distilled water to measure bound EPS. Soluble EPS, bound EPS and entire colloidal carbohydrates were measured by the phenol–sulfuric acid assay [12] with Dglucose dissolved in H2O used as a standard.

Scanning Electronic Microscopy (SEM) At the beginning of the experiment, small cores (5 mm in diameter and 5 mm in height) were filled with mud and placed with the top at the sediment surface. Three cores were removed from the mud at 11 AM (mid low-tide) on the 6th, 8th, and 10th days, glued onto a stub of the electronic microscope (JEOL 5410 LV, CCA, University of La Rochelle), and placed into liquid nitrogen haze (rather than plunged in liquid nitrogen, which could disturb natural structures) until it was transferred into the microscope chamber. At the 8th day, three cores were also taken at 7:30 AM (30 min before the beginning of daytime emersion period), at 8:30 AM (30 min after the beginning of the daytime emersion period), at 1:30 PM (30 min before the end of the daytime emersion period), and at 2:30 PM (30 min after immersion) to follow the migration pattern of diatoms during one experimental tide. Sediment cores were handled and stored as above. Uncoated sediment samples were placed in the microscope chamber, where the surface water gently sublimed off and the cold core was then gradually heated. The sediment samples were then removed from the microscope, coated with gold and replaced into the microscope chamber to be examined at a high acceleration voltage (between 10 and 15 kV). For each sample, six randomly chosen areas were photographed. With reference to recommendations by Paterson [23–25], we cautiously observed sediment surface during gradual sublimation and could check that no structure modification appeared due to dehydration. No broken cells were noted at the end of this stage and remnants of EPS mucilage matrix were not modified during dehydration compared to original structures. Preliminary observations were also made before this experiment and we checked that the employed method prevented artifacts that were obtained when using more rapid freeze-drying methods. Each picture was treated for the determination of the surface covered by diatoms relative to the whole surface. By lining up a transparent slide over every picture, each diatom which appeared on the picture was traced on the slide. Each slide was then scanned in a counting black-and-white bitmap format and pictures were finally analyzed using an algorithm (programmed in Visual Basic) for counting black and white pixels.

Statistical Analyses Comparisons between the fixed factors (i.e., depth, experimental tank, and experimental time) were made by analysis of variance

F. Orvain et al. using the MINITAB package, release 10.2. Our experimental setup was originally designed to perform ANOVA on ‘‘experimental tank,’’ ‘‘dark:light period’’ and ‘‘time’’ factors with an appropriate replication (n = 3). Unfortunately, no parametric tests could be used on chloropigment contents and colloidal carbohydrate contents due to heteroscedasticity (Bartlett’s test, P < 0.01) that was always found even when data were log(N+1) transformed. Consequently, for all pigment and carbohydrate data, we performed statistical tests on averaged net production data, which were obtained by computing the differences between averaged contents at both sampling times (i.e., at the end of light and dark periods). Day and night net production values were computed as follows: (biomass at 1:55 PM minus biomass at 8:05 AM) and (biomass at 8:05 AM minus biomass at 1:55 PM), respectively. In statistical tests, the replication (n = 2) finally consisted of averaged net production values computed for both tank data. Concerning changes in depth profiles of Chl a and pheopigments through ‘‘experimental time,’’ an appropriate way of testing ‘‘depth’’ effects was to perform MANOVA, because the same core was repeatedly measured along the depth and data from depth profiles would be spatially correlated [42]. Two separate MANOVA, which included two univariate one-way ANOVA to the ‘‘date’’ factor, were thus performed on each data block (i.e., day and night averaged net production data). All constraints for MANOVA were fully respected according to von Ende [42]. Concerning soluble EPS, bound EPS, and colloidal carbohydrates, measurements were done only on surface samples, and we performed two one-way ANOVA (including ‘‘date’’ factor) on averaged net production data on day and night data, separately. Concerning nutrient contents, we performed MANOVA to test the effects of ‘‘depth’’ factor due to the spatially correlated measures inside one core. MANOVA included two univariate two-way ANOVA that tested ‘‘date’’ and ‘‘day:night period’’ effects and their interaction.

Results Growth Observation Both replicate curves clearly showed that chlorophyll a contents increased following a sigmoid logistic-type curve in the uppermost centimeter compared to both replicate curves of control in the 1–2 cm layer (Fig. 2A). Three successive phases could be identified during this experiment: (1) a lag phase (ca. first 5 days), during which microphytobenthic cells adapted to the experimental tidal and diurnal forcings; (2) a logarithmic phase (from the 5th to the 9th day of the experiment), during which microphytobenthic cells were photosynthetically active and formed a biofilm; (3) a stationary phase, where the biomass reached its maximum on the 9th day and did not


241

Fig. 2. Pigment concentration kinetics for chlorophyll a (A) and pheopigments (B). Biomasses are plotted as a function of cultivation time (days) for the uppermost centimeter (black and gray triangles) and for the 1–2 cm layer (black and gray circles). Black rectangles represent immersion dark periods; daytime emersion periods are in between. Error bars represent standard deviations (n = 3).

increase for the remainder of the culture period. MANOVA performed on averaged net production revealed significant effects of ‘‘date’’ factor, but only on surface and during daytime (P £ 0.001) with difference in ‘‘date’’ effects between both 0–1 cm and 1–2 cm layers (P £ 0.05). Thus, it is clear that microphytobenthic production occurred during the experimental daytime emersion in the uppermost centimeter. The growth features reported from chlorophyll a kinetics were not found for pheopigments (Fig. 2B, MANOVA, P > 0.05). In our experiments, we removed macrofauna, but meiofauna was still present in the experimental sediment: foraminifera and nematoda were not abundant, with a total density equal to 85,000 ± 25,000 SD ind m)2 (n = 3), which could exert a limited grazing pressure on the microphytobenthic population. The water Chl a content

measured in the recuperation tank remained steady (data not shown) indicating that there were no diatom losses in the system. Ammonium content clearly decreased between the first and last days of the experiment within the surficial layer (Table 1, MANOVA, P £ 0.001). A significant influence of the depth was also obtained for ammonium contents (i.e., MANOVA, P £ 0.005) and this indicates that the stationary phase was probably initiated in response to an ammonium depletion. No significant variation occurred between depths or dates for other nutrients (P > 0.05). Carbohydrate Kinetics A strong correlation was found between colloidal carbohydrate contents and the sum of soluble and bound EPS

242

F. Orvain et al.

Table 1. Nutrient concentrations (lM) with their standard deviations at the beginning and the end of the experiment Experimental conditions Sampling day

Sampling time

First day

8.05 AM 1.55 PM

Last day

8.05 AM 1.55 PM

Depth 0–1 1–2 0–1 1–2 0–1 1–2 0–1 1–2

cm cm cm cm cm cm cm cm

Urea 11.1 12.0 10.8 9.6 5.5 8.2 17.8 14.2

± ± ± ± ± ± ± ±

fractions (r = 0.92, n = 149, P < 0.001) with a slope close to 1 (i.e., 1.03 ± 0.07 SD). This allows us to confirm the reliability of carbohydrate measurements. There were significant differences among experimental times for colloidal carbohydrates as well as bound EPS productions (Figs. 3A and B, one-way ANOVA; P < 0.001). In contrast, soluble EPS net production was constant with time (one-way ANOVA, P > 0.05; Fig. 3C). Bound EPS contents increased rapidly during the light periods over the stationary phase (Fig. 3B). These increases were followed by drastic falls during night submerged periods, corresponding to an almost complete loss of material. As a result, this led to an overall jagged pattern of bound EPS content kinetics. The range over which bound EPS contents fluctuated was less high during the logarithmic phase (Fig. 3B). The same daily fluctuations were observed for soluble EPS contents over the course of the experiment (Fig. 3C) but, unlike bound EPS, no enhancement of diurnally soluble EPS net growth rates was observed once cultures entered the stationary phase. Approximately 60% of colloidal carbohydrate contents were represented by bound EPS over the course of the experiment. As a result, colloidal carbohydrate contents (Fig. 3A) showed oscillations, which were comparable to the bound EPS kinetics pattern, with more pronounced productions during stationary phase compared to logarithmic phase.

Microphytobenthic Migratory Processes Natural benthic populations taken from the Aiguillon Cove sediments were used for our experiments and diatom taxa were identified by SEM. Dominant species or genera were Gyrosigma fasciola, Navicula spp., Nitzschia spp., Pleurosigma angulatum, P. formosum, and Pectrodictyon gemma (Figs. 4 and 5).

2.1 1.7 6.8 1.8 1.4 0.5 0.0 4.8

Ammonium 330.4 552.7 348.2 606.6 14.5 302.0 61.6 462.5

± ± ± ± ± ± ± ±

29.7 38.8 109.6 48.7 7.1 376.2 15.1 94.0

Nitrate Nitrite 13.3 11.1 9.1 12.3 22.1 11.0 11.0 13.1

± ± ± ± ± ± ± ±

8.0 3.2 2.7 2.9 14.9 0.0 5.5 3.1

Phosphate 2.8 3.3 2.7 1.6 7.2 5.3 5.1 3.6

± ± ± ± ± ± ± ±

0.4 2.2 1.5 0.5 6.0 0.2 2.7 1.5

A high variability between cores partially masked trends in migration patterns with regard to the scale observation of SEM and patchy distribution of diatoms shown in Fig. 5. Spatial heterogeneity in microphytobenthic biofilms is typical from microphytobenthic assemblages. Biofilms are acknowledged to form patches at the beginning of the logarithmic period, with patch sizes growing until the sediment surface was totally covered. The small cores (5 mm in diameter), which were taken randomly, could be taken either from sediment covered by a patchy biofilm, or from between-patches sediments not covered by a biofilm (Fig. 5). Despite the spatial heterogeneity, pictures from the 8th day of the experiment (Figs. 4 and 6A) indicated the occurrence of migration patterns. Mean percentages of sediment surface covered by diatoms were maximal at the mid-exposure time and minimal values were found at both water cover periods (Fig. 6). At the end of the experimental high tide, few diatoms were present at the water–sediment interface (Fig. 4A) and no EPS filaments appeared. At the beginning of the exposure period, diatoms began to migrate toward the sediment surface (Fig. 4B). In the middle of the daytime air exposure period, the biofilm was clearly present (Fig. 4C). At the end of the daytime air exposure period (Fig. 4D), diatoms again became scarce, but the development of the embedding EPS matrix was maximal. Thirty minutes after the artificial tide rose (Fig. 4E), fibrous dehydrated EPS in association with aggregated material could still be observed on pictures in spite of the water cover. During the artificial high tide (Figs. 4A and 4E), only big species (Gyrosigma fasciola, Pleurosigma angulatum) could be observed at the surface of sediment, whereas high densities of small species (Navicula spp., Nitzschia spp.) were observed on pictures taken during emersion periods (Figs. 4B, 4C, and 4D).


Fig. 3. Carbohydrate concentration kinetics in mesocosm for colloidal carbohydrates (A), bound EPS (B), and soluble EPS (C). Biomasses are plotted as a function of cultivation time (days) for the uppermost centimeter (black and gray triangles). Black rectangles represent immersion dark periods; daytime emersion periods are in between. Error bars represent standard deviations (n = 3).

Relation between Sediment Carbohydrate Fractions and Chl a Positive correlations were found between Chl a biomass and all carbohydrate fractions. The correlation with soluble EPS was found to be the strongest (r = 0.80, n = 42, P < 0.001, compared to r = 0.64, n = 42, P < 0.001 and

243

r = 0.60, n = 42, P < 0.001 for colloidal carbohydrates and bound EPS fraction, respectively). No relationship was found between phaeopigment biomass and the three carbohydrate fractions (r = 0.11, n = 42, P > 0.05; r = 0.14, n = 42, P > 0.05; r = 0.31; n = 42, P > 0.05 for soluble EPS, bound EPS, and colloidal carbohydrates, respectively). For the logarithmic phase, diurnal net production rates were 0.35 lg g sediment)1 h)1 and 2.02 lg g sediment)1 h)1 for soluble EPS and bound EPS, respectively. For the stationary phase, diurnal net production rates were equal to 19.79 lg g sediment)1 h)1 and 4.66 lg g sediment)1 h)1 for soluble EPS and bound EPS, respectively. Thus, approximately 15% of the diurnal net carbohydrate production consisted of bound EPS during the logarithmic phase, and this proportion reached approximately 80% during the stationary phase. For both fractions, production rates were maximum as cultures moved from logarithmic growth into the stationary phase. These values were equal to 1.23 lg lg Chl a)1 h)1 and 0.44 lg lg Chl a)1 h)1 for bound EPS and soluble EPS, respectively. There was a significant linear increase (r2 = 0.81, n = 16, P < 0.001; Fig. 7A) between produced EPS and produced Chl a during the logarithmic phase. The four values from the stationary phase deviated from this linear relationship and bound EPS production was exacerbated during this period, whereas Chl a biomass reached its biotic capacity. This suggests another relationship existing between bound EPS and Chl a production during the stationary phase. During the logarithmic phase, production of soluble EPS asymptotically increased with Chl a production and leveled off ca 6 lg g sediment)1 h)1. This relationship led us to test a von Bertalanffy equation (r2 = 0.68, n = 16, P < 0.001; Fig. 7B). Another relationship may be occurred during the stationary phase. However, only one point (exceeding the plateau obtained during the logarithmic phase) clearly deviated from exponential phase relationship and no further assumptions can be based on this observation. As a result of both fractions, colloidal carbohydrates increased linearly versus produced Chl a during daytime air exposure periods of the logarithmic phase (r2 = 0.89, n = 16, P < 0.001; Fig. 7C). A change in diatom carbohydrate metabolism occurred during the stationary phase compared to the logarithmic phase, since the four values of produced carbohydrates calculated for the stationary

244

F. Orvain et al.

Fig. 4. SEM of sediment surfaces (bar marker 50 lm) taken from mesocosm on the 8th day of culture. (A) At the end of artificial hightide (i.e., at 7:30 AM), a low diatom population was present on the sediment surface and sediment particles appeared unconnected. (B) At the beginning of the daytime air exposure period (i.e., at 8:30 AM), more diatoms were present at the sediment surface than before and an extensively developed extracellular matrix appeared. (C) In the middle of the artificial daytime air exposure (i.e., at 11 AM), diatoms were even more abundant. (D) At the end of the exposure period (i.e., at 1:30 PM), diatoms were less abundant than at mid-exposure time, but the development of EPS matrix was maximal. (E) At the beginning of the artificial hightide (i.e., at 2:30 PM), diatoms were scarce again, but individual particles showed a network of delicate EPS strands linking them with surrounding structures. Note that only large species were present on pictures corresponding to immersion periods (A and E).

phase also deviated from the relationship found in the logarithmic phase.

Discussion Simulating In Situ Microphytobenthic Biofilm Development

Moisture Content Analyses Sediment moisture content increased during the experiment and followed a sigmoid logistic-type curve in the uppermost centimeter, unlike control in the underlying layers. During the logarithmic phase, the sediment moisture content converged toward a maximal value of 171.4%, which remained stable during the stationary phase. A positive correlation was found between sediment moisture content and Chl a content (r = 0.78, n = 18, P < 0.01). There was also an increase in all the carbohydrate extracts with increasing sediment moisture content in all pooled data with significant relationships (r = 0.74, n = 18, P < 0.001; r = 0.60, n = 18, P < 0.01; r = 0.70, n = 18, P < 0.01 for LMW colloidal carbohydrates, EPS and total colloidal carbohydrates, respectively).

Diatom losses, due to natural resuspension and grazing, were prevented in the mesocosm, and microphytobenthos biomass thus followed a logistic-type curve by converging toward a maximal value termed ‘‘the biotic capacity of the local environment’’ (Fig. 2A). At a daily scale, a significant increase in microphytobenthic biomass was only observed during daytime emersion periods. SEM pictures have often been used to study diatom migratory rhythms typical of natural intertidal situations [1, 16, 17, 19, 23–26, 31, 43]. For instance, Paterson [24] reported that ‘‘in the initial stages of the low-tide, EPS strands were associated with diatoms emerging at the sediment surface, but soon increased in cover and density. At the end of the low-tide periods, the EPS matrix persisted once the diatoms had left the surface until the return of the tide, where a significant lessening in the extent of


245

Fig. 5. SEM of sediments (bar marker 50 lm) taken from mesocosm in the middle of the artificial daytime air exposure (i.e., at 11 AM) on the 10th day of culture. Two pictures are shown from each of the three cores (A, B, and C) to account for the very high spatial variability.

cover was noted.’’ Other techniques have involved the use of the PAM fluorometer, which has yielded insight into the examination of diatom migratory patterns [27, 28, 30]. We chose to use SEM in combination with picture analyses to assess the occurrence of migratory rhythms in our mesocosm experiments and quantify them. Features mentioned above for in situ migratory rhythms were all reproduced within one tidal cycle (Figs. 4 and 6A) leading to the diatom development (Fig. 2A). The acknowledged spatial heterogeneity in biofilm was also observed on SEM (Figs. 5 and 6) and also appeared through standard deviations of Chl a biomass during exponential phase (error bars on Fig. 2A). Many laboratory studies have reported that diatom growth follows a logarithmic pattern before reaching a maximal value, which acts like a biotic capacity during the stationary phase. This classical logistic development has been reported in diatom cultures for both water column [1, 7, 29, 34, 33, 40] and sediment interface [36]. Generally

in intertidal mudflats, grazing and resuspension cause biomass losses due to which diatoms remain in the logarithmic phase of their development. This was not the case in our mesocosm, in which we removed the processes responsible for biomass losses. We can thus experiment with the full development of a biofilm whose different stages can be observed in the field as a result of environmental conditions and the age of the biofilm. Indeed, our mesocosm simulates natural situations similar to the natural environment during the logarithmic phase, and also simulates natural situations of the stationary phase, which can occasionally occur in natural conditions [43, 44]. Short-Term Variations in Carbohydrate Productions We found that the strongest relationship between Chl a biomasses and carbohydrates was obtained with soluble EPS in comparison to colloidal carbohydrates and bound

246

F. Orvain et al.

Fig. 6. Averaged percentages of sediment covered by diatoms SEM (crosses) with their standard deviations versus the sampling times on the 8th day of the experiment. Percentages obtained from each sample (circles) show the spatial variability of the biofilm.

EPS fraction. This is also in accordance with other in situ studies [6, 14, 26, 39, 41, 44]. In our experiments, Chl a, soluble and bound EPS kinetics were clearly different (Figs. 2 and 3), but Chl a and carbohydrate production always occurred during daytime emersion periods. Our data thus support the idea that secretion of EPS is dependent on photosynthesis patterns [6, 7, 31, 35]. Rince´ et al. [29] demonstrated that the coverage of a substrate by EPS matrix occurred before Chl a began to increase, suggesting that at least a small amount of produced EPS is required by diatoms during their locomotion at the beginning of the logarithmic phase. Since Chl a and EPS diurnal net productions were positively and linearly correlated in the mesocosm and since we have indication that migratory features occurred in a classic way, our data agree with the theory according to which the synthesis of EPS is implicated in attachment and locomotion of diatoms, which is a necessary pattern for diatoms to realize their diurnal photosynthesis [28, 31, 32]. Nevertheless, diatom motility and associated EPS secretion are expected to occur during night periods as well as light periods [31, 32]. In the mesocosm, EPS secreted during night immersions were probably rapidly washed away and we cannot state anything from our results about night EPS production that could be related to diatom locomotion. Drastic falls were found in EPS contents during night immersion of the stationary phase and, to a lesser extent, decreases were found in all three carbohydrate fractions during night periods whatever the phase. These disappearances of EPS during night immersions were also observed both in culture as in the field by Underwood and Smith [40], van Duyl et al. [41], Staats et al. [35], and de

Fig. 7. Both diurnal produced bound EPS (A) and soluble EPS (B) vs the diurnally produced Chl a during the logarithmic (solid symbols) and the stationary phases (open symbols). Regression lines of equations [produced bound EPS] = 16.06 · [produced Chl a], r2 = 0.81; [produced soluble EPS] = 5.80 · (1)e)4.29·[produced chl a]), r2 = 0.68 and [produced colloidal carbohydrate] = 33.35 · [produced Chl a], r2 = 0.89 were calculated using logarithmic phase data.

Brouwer and Stal [6]. Many hypotheses have been proposed by these authors to account for these decreases: (1) The extracellular carbohydrates were washed away and dissolved into the water column when the tide rose [6, 10, 35, 40]; (2) the presence of microbial enzymes such as bglucosidase secreted by bacteria [15, 41] could accelerate polysaccharide hydrolyses; and (3) diatoms themselves are able to hydrolyze their own extracellular polysaccharides


and hydrolyzed products could be taken up by diatoms, which could act as facultative heterotrophs [7, 35]. De Brouwer and Stal [7] have found that resorption of EPS by diatoms provided the best explanation for bound EPS decrease during night periods in axenic cultures because polymeric bound EPS fractions disappeared in their cultures without entailing a concomitant increase in smaller sugar chains of soluble EPS. They also demonstrated that both fractions were always unconnected in term of sugar composition, providing another argument in favor of the diatom heterotrophy assumption rather than EPS dissolution. In the mesocosm, soluble EPS as well as bound EPS disappeared during night immersion periods. This result suggests that hydrolysis and/or washout affected both EPS fractions similarly. The major part of bound EPS was certainly dissolved in the seawater and transferred to soluble EPS which were hydrolyzed in turn. EPS degradation must also be under the control of enzymatic degradation, and bacteria as well as diatoms were likely to be implicated in these processes. The three assumptions mentioned to interpret soluble and bound EPS decreases were therefore likely to interact in mesocosm where complex benthic interactions were reproduced. However, bacterial degradation of EPS is less likely to occur because van Duyl et al. [41] reported that the drastic fall in soluble EPS is not attributable to bacteria during submerged periods, but rather a high amount of soluble EPS is bioavailable for bacteria at the end of emerged periods, when bacterial production is at its highest within the tidal cycle. Unlike de Brouwer and Stal, who found a continuous release of soluble EPS in axenic medium monocultures [7], soluble EPS contents showed oscillation in the mesocosm in response to the alternating day:night cycle. In our case, daily production of soluble EPS may be related to endogenous migratory rhythms, which governed microphytobenthic activity in mesocosms [25], while endogenous rhythms cannot be induced in medium cultures, where diatom locomotion processes must probably be constant over time. Long-Term Variations in Carbohydrate Production For bound EPS, there was a clear opposition between logarithmic and stationary phase situations. During the logarithmic phase, there was a positive relationship between bound EPS produced and diatom biomass produced (Figs. 3B and 7A) while, during stationary phase, bound EPS diurnal production became very high and Chl a

247

produced became very low. This enhancement in bound EPS production confirms the theory, according to which EPS exudation is a function of the physiological state of algae related to nutrient source depletion [1, 10, 33]. Nutrient concentration differences between the beginning and the end of the experiment suggest that ammonium was depleted. Probably in response to this mechanism, a modification of algal metabolism occurred when cells entered the stationary phase, since ca. 80% of total carbohydrates were bound EPS during the stationary phase compared to 20% during the logarithmic phase. The range of increase in bound EPS proportions during stationary phase agrees with culture and field measurements [10, 31, 32, 36, 40, 44]. When cells entered the stationary phase, our data suggest a shift in metabolic pathway allowing diatoms to continue photosynthesis in unfavorable conditions (nutrient limitation) by secreting photosynthetic products through overflow metabolism. This modification in metabolism did not similarly affect soluble EPS production, for which no opposition between logarithmic and stationary phase situations was clearly detected (Figs. 3C and 7B). Our results are in conformity with those of de Brouwer and Stal [7] who demonstrated that soluble and bound EPS are regulated by different metabolic controls. However, they reported a different situation from us since the highest bound EPS production occurred during logarithmic phase in their study. These divergent results can be interpreted by different causes of stationary phase emergence between the two distinct studies. Indeed, we have indications that stationary phase was induced by nutrient limitation and de Brouwer and Stal did not attribute the stationary phase to that mechanism, but rather to a light limitation due to self-shading that occurred in their medium cultures. Since EPS production has been demonstrated to be strictly light dependent [6, 8, 32, 35], a decrease in bound EPS production occurred in response to the decrease in light intensity. This comparison between studies illustrates the straight dependence of bound EPS production on environmental conditions. Soluble EPS could be simple sugars comprising galactose, mannose, ribose, xylose, fucose, rhamnose, and mainly glucose[6–8, 10, 34, 37]. We can suggest that these intracellular products provide the main photoassimilates available in cells for mucilage secretion, which is necessary for locomotion during the first stages of the logarithmic phase. Indeed, the main carbohydrates secreted during the first stages of the logarithmic phase were low molecular

248

F. Orvain et al.

Fig. 8. Sediment moisture content kinetics in the mesocosm. Sediment moisture contents are plotted as a function of time (days) for the uppermost centimeter (triangles) and for the 1–2 cm layer (circles). Black rectangles represent immersion dark periods; daytime emersion periods are in between. Error bars represent standard deviations (n = 4).

weight compounds (i.e., soluble EPS). The continuous daytime release of soluble EPS over the full growth curve (Fig. 3C) led us to relate soluble EPS production to diatom locomotion related to migratory rhythms. Carbohydrates that are exuded by marine benthic diatoms form a hydrated matrix of strands embedding diatoms and sediment particles [16, 23, 25] and it has been demonstrated from medium cultures that soluble EPS represent a fraction that is not or is only loosely associated with diatom cells, while bound EPS are associated with diatom aggregates [8]. As a result, bound EPS may remain associated with cells even once diatoms leave the sediment–water interface during immersion periods, and these compounds could be further available for diatoms capable of facultative heterotrophy as suggested by de Brouwer and Stal [7, 8]. On the other hand, soluble EPS must remain at the sediment–water interface during water coverage and must be included in the EPS strands that were observed during immersion periods on SEM (Fig. 4E). Soluble EPS could then be a better indicator of locomotion processes than bound EPS since they behave like remnants of diatom biofilm during their movement. During the logarithmic phase, there was a clear positive relationship between produced colloidal carbohydrate and produced Chl a during daytime emersion periods (Fig. 7C). The positive relationships with both soluble and bound EPS fractions were not so strong as for entire colloidal carbohydrates and were clearly nonlinear for soluble EPS. We found a maximum of soluble EPS production during the logarithmic phase leading to a von Bertalanffy type curve versus Chl a production. Two hypotheses can account for the nonlinearity of this relationship. First, at the end of the logarithmic phase, the decrease in Chl a

production when reaching the biotic capacity was first initiated by a limitation in production rates of soluble EPS. Second, the development of bacterial populations could provide an alternative explanation for the reduction in soluble EPS production during the final stages of the logarithmic phase. Bacteria could develop on high amounts of soluble EPS when diatoms entered the stationary phase [44]. An increase in the bacterial population could be related to a decrease in soluble EPS, which could provide directly assimilable and bioavailable food at the end of emerged periods [41]. Unfortunately, bacteria were not evaluated during our experiments and therefore it is difficult to check this assumption. Potential Influences of Biofilm Development on Sediment Stability Various possibilities can be considered that describe the potential influence of EPS amounts on sediment stability, and this highlights the difficulties in understanding biological effects on sediment properties due to complex benthic interactions. For instance, during daytime emersion periods, diatoms secreted EPS and these microbial secretions might affect sediment stability through binding processes [3, 17, 22, 36, 38]. Colloidal carbohydrate production increased with Chl a production over the full growth curve and carbohydrate contents (at the start of immersion periods) were maximal during stationary phase. As a result, the degree of EPS influence on sediment stability was likely to increase over the rapid growth of diatom biofilm. However, the potential stabilizing influence must be limited to a very tight temporal window, when the tide rose during flood tide, since carbohydrates disappeared when submerged in the mesocosm (Fig. 3).


Since no EPS buildup can accumulate over time, we may eventually suggest that EPS are not involved in sediment stabilization. We also observed an increase in sediment moisture content with the increase in diatom biomass (Fig. 8) due to EPS secretions, which could act like a spongy structure and limit water exchange at the interface [11]. A major function of EPS coatings may be to conserve water during periods of tidal exposure [11]. A similar positive influence of diatoms on sediment moisture content was also reported from the field [26] and another mesocosm study [44]. The latter authors suggest that water is directly associated with the EPS matrix and there may be relatively little interstitial water. Traditionally, there is a general consensus among sedimentologists that moisture content is the main variable controlling cohesive sediment stability and many authors have demonstrated that the wettest sediments are less able to resist tidal shear stresses. In this context, we can expect that sediment beds cultivated in the mesocosm were less stable during the stationary phase compared to the logarithmic phase, even though they were the most EPS-enriched during stationary phase. This argument points out that there are still paradoxes in elucidating microphytobenthic actions on sediment stability and we can still ask questions about the best indicator for sediment stability, which can be moisture content as well as EPS content. According to the chosen criteria, opposite conclusions can be drawn from our data. Furthermore, the extent to which different EPS fractions are implicated in sediment stabilization is unclear and must be further investigated.

249

dark cycles. Our overall findings are in agreements with axenic medium monocultures [7, 33] and confirmed that soluble EPS and bound EPS fractions exhibited different kinetics depending on environmental conditions. Although our data do not allow further determination of effective influences of microphytobenthic biofilm development on sediment stability, they illustrate the necessity to carry out new experiments to assess the extent to which the distinct EPS fractions are implicated in sediment stability.

Acknowledgments The Regional Council of Poitou-Charentes and Ifremer have financially supported the study by a doctoral grant awarded to F.O. We are grateful to D. Gouleau, M. Laborde, R. Boutin, and A. Sygut for their help and advice in the production of electronic microscopy pictures, P. Richard for his help in the analyses of pictures, and F. Mornet for her help in the determination of diatom species. Many thanks to the three referees for all their reliable suggestions helping us to improve the manuscript.

References 1. 2.

Conclusion We suggest that our integrated approach, which encompassed the accurate description of biomass dynamics behavior, carbohydrate production, and migration patterns, all related to biofilm development, could provide a system of reference for new in situ studies. For instance, the characterization of the logarithmic and stationary phases in the natural environment could be carried out by measuring all criteria of the biofilm development studied here. This integrated approach in mesocosm allowed us to identify several properties of biofilm development, which emerged separately from previous laboratory and in situ studies, such as the daily oscillations of bound EPS as well as soluble EPS in response to tidal and alternating light:

3.

4.

5.

6.

7.

Admiraal W (1984) The ecology of estuarine sediment-inhabiting diatoms. Prog Phycol Res 3:269–322 Aminot A, Chaussepied M (1998) Manuel des analyses chimiques en milieu marin. CNEXO, Brest Andersen TJ (2001) Seasonal variations in erodability of two temperate, microtidal mudflats. Estuar Coast Shelf Sci 53:1– 12 Blanchard GF, Cariou-Le Gall V (1994) Photosynthetic characteristics of microphytobenthos in Marennes-Ole´ron Bay, France: preliminary results. J Exp Mar Biol Ecol 182:1– 14 Blanchard GF, Guarini JM, Orvain F, Sauriau PG (2001) Dynamic behaviour of benthic microalgal biomass in intertidal mudflats. J Exp Mar Biol Ecol 264:85–100 de Brouwer JFC, Stal LJ (2001) Short-term dynamics in microphytobenthos distribution and associated extracellular carbohydrates in surface sediments of an intertidal mudflat. Mar Ecol Prog Ser 218:33–44 de Brouwer JFC, Stal LJ (2002) Daily fluctuations of exopolymers in cultures of the benthic diatoms Cylindrotheca closterium and Nitzchia sp. (Bacillariophyceae). J Phycol 38:464–472

250 8.

9.

10.

11. 12.

13. 14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

F. Orvain et al. de Brouwer JFC, Wolfstein K, Stal LJ (2002) Physical characterization and diel dynamics of different fractions of extracellular polysaccharides in an axenic culture of a benthic diatom. Eur J Phycol 37:37–44 Cariou-Le Gall V, Blanchard GF (1995) Monthly HPLC measurements of pigment concentration from an intertidal muddy sediment of Marennes-Ole´ron Bay, France. Mar Ecol Prog Ser 121:171–179 Decho AW (1990) Microbial exopolymer secretions in ocean environments: their role(s) in food webs and marine processes. Oceanogr Mar Biol Annu Rev 28:73–153 Decho AW (2000) Microbial biofilms in intertidal systems: an overview. Cont Shelf Res 20:1257–1273 Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F (1956) Colorimetric method for determination of sugars and related substances. Anal Chem 28:350–356 Edgar LA, Pickett-Heaps JD (1984) Diatom locomotion. Prog Phycol Res 3:47–88 Galois R, Blanchard G, Seguignes M, Huet V, Joassard L (2000) Spatial distribution of sediment particulate organic matter on two estuarine mudflats: a comparison between Marennes-Ole´ron Bay (France) and the Humber Estuary (UK). Cont Shelf Res 20:1199–1217 Goto N, Mitamura O, Terai H (2001) Biodegradation of photosynthetically produced extracellular organic carbon from intertidal benthic algae. J Exp Mar Biol Ecol 257:73– 86 Gouleau D (1976) Le roˆ le des diatome´es benthiques dans l’engraissement rapide des vasie`res atlantiques de´ouvrantes. CR Acad Sci Paris 283:21–26 Grant J, Bathmann UV, Mills EL (1986) The interaction between benthic diatom films and sediment transport. Estuar Coast Shelf Sci 23:225–238 Guarini JM, Blanchard GF, Gros P, Gouleau D, Bacher C (2000) Dynamic model of the short-term variability of microphytobenthos biomass on temperate intertidal mudflats. Mar Ecol Prog Ser 195:291–303 Happey-Wood CM, Jones P (1988) Rhythms of vertical migration and motility in intertidal benthic diatoms with particular reference to Pleurosigma angulatum. Diat Res 3:83–93 Kelly JA, Honeywill C, Paterson DM (2001) Microscale analysis of chlorophyll-a in cohesive, intertidal sediments: the implications of microphytobenthos distribution. J Mar Biol Ass UK 81:151–162 Lorenzen CJ (1966) A method for the continuous measurement of in vivo chl concentration. Deep-Sea Res 13:223– 227 Madsen KN, Nilsson P, Su¨ ndback K (1993) The influence of benthic microalgae on the stability of a subtidal shallow water sediment. J Exp Mar Biol Ecol 170:159–177 Paterson DM (1986) The migratory behaviour of diatom assemblages in a laboratory tidal micro-ecosystem examined by low temperature scanning electron microscopy. Diat Res 2:227–239

24. Paterson DM (1989) Short-term changes in the erodability of intertidal cohesive sediments related to the migratory behaviour of epipelic diatoms. Limnol Oceanogr 34:223–234 25. Paterson DM, Crawford RM, Little C (1986) The structure of benthic diatom assemblages: a preliminary account of the use and evaluation of low-temperature scanning electron microscopy. J Exp Mar Biol Ecol 96:279–289 26. Paterson DM, Tolhurst TJ, Kelly JA, Honeywill C, de Deckere V, Huet V, Shayler SA, Black KS, de Brouwer J, Davidson I (2000) Variations in sediment properties, across the Skeffling mudflat, Humber Estuary, UK. Cont Shelf Res 20:1373– 1396 27. Perkins RG, Oxborough, Hanlon ARM, Underwood GJC, Baker NR (2002) Can chlorophyll fluorescence be used to estimate the rate of photosynthetic electron transport within microphytobenthic biofilms? Mar Ecol Prog Ser 228:47–56 28. Perkins RG, Underwood GJC, Brotas V, Snow V, Jesus B, Ribeiro L (2001) Responses of microphytobenthos to light: primary production and carbohydrate allocation over an emersion period. Mar Ecol Prog ser 223:101–112 29. Rince´ Y, Lebeau T, Robert JM (1999) Artificial cell-immobilization: a model simulating immobilization in natural environments? J Appl Phycol 11:263–272 30. Seroˆ dio J, da Silva JM, Catarino F (1997) Non destructive tracing of migratory rhythms of intertidal benthic microalgae using in vivo chlorophyll a fluorescence. J Phycol 33:542–553 31. Smith DJ, Underwood GJC (1998) Exopolymer production by intertidal epipelic diatoms. Limnol Oceanogr 43:1578– 1591 32. Smith DJ, Underwood GJC (2000) The production of extracellular carbohydrates by estuarine benthic diatoms, the effects of growth phase and light and dark treatment. J Phycol 36:321–333 33. Staats N, Stal LJ, Mur LR (2000) Exopolysaccharide production by the epipelic diatom Cylindrotheca closterium: effects of nutrient conditions. J Exp Mar Biol Ecol 249:13–27 34. Staats N, de Winder B, Stal LJ, Mur LR (1999) Isolation and characterization of extracellular polysaccharides from the epipelic diatoms Cylindrotheca closterium and Navicula salinarum. Eur J Phycol 34:161–169 35. Staats N, de Winder B, Stal LJ, Mur LR (2000) Oxygenic photosynthesis as driving process in exopolysaccharide production of benthic diatoms. Mar Ecol Prog Ser 193:261– 269 36. Sutherland TF, Grant J, Amos CL (1998) The effect of carbohydrate production by the diatom Nitzschia curvilineata on the erodability of sediment. Limnol Oceanogr 43:65–72 37. Taylor IS, Paterson DM, Mehlert A (1999) The quantitative variability and monosaccharide composition of sediment carbohydrates associated with intertidal diatom assemblages. Biogeochemistry 45:303–327 38. Underwood GJC, Paterson DM (1993) Seasonal changes in diatom biomass, sediment stability and biogenic stabilization in the Severn Estuary. J Mar Biol Ass UK 73:871–887

Development of a Microphytobenthic Biofilm 39. Underwood GJC, Paterson DM, Parkes RJ (1995) The measurement of microbial carbohydrate exopolymers from intertidal sediments. Limnol Oceanogr 40:1243–1253 40. Underwood GJC, Smith DJ (1998) Predicting epipelic diatom exopolymer concentrations in intertidal sediments from sediment chlorophyll a. Microb Ecol 35:116–125 41. van Duyl FC, de Winder B, Kop AJ, Wollenzien U (1999) Tidal coupling between carbohydrate concentrations and bacterial activities in diatom-inhabited intertidal mudflats. Mar Ecol Prog Ser 191:19–32 42. von Ende CN (1993) Repeated measures analysis: growth and other time-dependent measures. In: SM Scheiner, J

251 Gurevitch (eds) Design and Analysis of Ecological Experiments. Chapman & Hall, London pp 113–137 43. Yallop ML, de Winder B, Paterson DM, Stal LJ (1994) Comparative structure, primary production and biogenic stabilization of cohesive and non-cohesive sediments inhabited by microphytobenthos. Estuar Coast Shelf Sci 39:565–582 44. Yallop ML, Paterson DM, Wellsbury P (2000) Interrelationships between rates of microbial production, exopolymer production, microbial biomass, and sediment stability in biofilms of intertidal sediments. Microb Ecol 39:116– 227