Exopolymer microenvironments of microbial flora: Multiple and interactive effects on trophic relationships

Limnol. Oceanogr., 38(R), 1993, 1633-1645 8 1993, by the American Society of Limnology and Oceanography. Inc

Exopolymer microenvironments of microbial flora: Multiple and interactive effects on trophic relationships

Alan W. Decho U.S. Geological Survey, 345 Middlefield Rd., MS 465, Menlo Park, California 94025

Glenn R. Lopez 3

Marine Sciences Research Center, SUNY, Stony Brook, New York 11794-5000

Abstract Microbial cells in natural environments arc often encased in different types of exopolymcr secretions

(EPS), ranging from tight capsules surrounding individual cells to the looser slime matrices of biofilms. The diffcrcnt physical and chemical propertics of exopolymers could have secondary effects on trophic interactions between microbial cells and consumer animals. Laboratory studies showed that capsule EPS is significantly less digestible to consumers than slime EPS, even when extracted from the same bacterial strain. Bacterial cells with EPS capsules are less efficiently digested than noncapsuled cells, suggesting that capsules protect against digestion. Follow-up experiments determined that polysaccharide-rich fractions of slime EPS are absorbed with very high eficiencies while protein portions, which are more abundant in capsular polymers, are absorbed relatively poorly. Another series of cxpcriments showed that dissolved organic matter (DOM), when adsorbed directly to the mineralogical portions of sediment particles, is available to deposit feeders. However, the further presence of an exopolymer coating on sediments more than doubled the bioavailability of adsorbed DOM to the consumer. Observations using cold-stage scanning electron microscopy indicated that exopolymer microenvironments are a common feature of natural marine sediments. Microbial exopolymers range from easily digestible carbon sources to relatively refractory ones that effectively protect some microbial cells from consumer digestion. Exopolymer mi- crocnvironments may also make recently adsorbed DOM highly accessible to particle-ingesting animals.

Exopolymers are large-molecular-weight (> 100,000 daltons) mucous secretions of microbial cells. The physiological state of the microbial cell and local environmental conditions often dictate the type and composition of polymers that are secreted and for what functions (Sutherland 1977). The same micro- organism, for example, can secrete several different exopolymers simultaneously or in se- quence (Christensen et al. 1985). Differences in the composition and physical properties of exopolymer secretions (EPS) also may affect their digestibility to animals that ingest the microbial flora associated with sediments and water-column aggregates. The few feeding studies conducted thus far examining use of EPS by animals (Baird and Thistle 1986; De- cho and Moriarty 1990) have indicated that some exopolymers represent highly labile food

Acknowledgments We thank R. Petty for C: N analyses of exopolymer

samples and R. Oscarson for technical assistance in cold- stage SEM preparations.

This research was supported by an HRF grant (012: 89A123) and the USGS.

sources. However, based on the potential of exopolymers for having differing compositions and properties, there is reason to suspect a priori that some of them may be relatively refractory.

Exopolymers can be classified into two major types based on their general physical structure: capsule exopolymers, which closely sur- round cells, and slime exopolymers, which are loosely associated with cells. Slime exopolymers are often largely polysaccharide in composition and show a much looser tertiary con- figuration than capsule polymers (Rees et al. 1982). They can exist as particulate organic matter (POM) in aggregates or biofilms on the surface of particles or as dissolved organic matter (DOM) in solution. Capsules perform important functions which enhance the sur- vival and propagation of microbial cells under natural conditions (see Decho 1990). They are more tightly wound in their physical configu- ration. Some capsule exopolymers are typically higher in glycoproteins than their slime counterparts (Sutherland 1977). Such differences between capsule and slime exopolymers may be of trophic importance, because these

1633

1634 Decho and Lopez

properties can affect the digestibility of the polymer or even microbial cells within the polymer.

Selective digestion and the ability of certain microbial cells to resist digestion by consumer animals have been frequently noted (Decho and Castenholz 1986 and citations therein). Bacterial and microalgal cells often remain intact and viable after passage through a consumer gut. Some microalgae have even been shown to photosynthesize (Epp and Lewis 198 1) and obtain nutrients (Porter 1976) during such passage. It has been postulated that the ability of some microbial cells to resist digestion is due to differences in cell-wall composition and perhaps the presence of gelati- nous capsular coatings (Porter 1976). Capsule EPS may slow the access of hydrolyzing enzymes to the microbial cell during digestion, thus increasing the cell’s chances of surviving gut passage. Therefore, while some slime EPS represent easily digestible foods for consumers, capsule EPS often show properties that might render them more refractory.

Particulate organic carbon (POC), such as cells, sediment-associated detritus, and amorphous material, is efficiently ingested by animals. DOM, however, either as solute or colloidal compounds, is not readily available to most animals (see Gomme 1982). It has been generally recognized that the binding of dissolved compounds to particles can potentially make the dissolved compounds accessible to particle-ingesting animals. However, few em- pirical studies have directly addressed this idea or the potential mechanisms which govern the accessibility of sediment-bound DOM to consumers.

Exopolymers may contribute to the accessibility of DOM to consumer animals. The EPS matrix represents a specific microenvironment for the adsorption of different types of DOM. Exopolymers contain abundant adsorptive ligands that sequester DOM. This adsorbed DOM provides a reservoir of nutrients and organic compounds which can be hydrolyzed to monomers and taken up and metabolized by microbial cells within the matrix. Biofilm coatings on sediments, aggregates, and other particles contain an abundance of EPS. Sedi- ment-dwelling animals which ingest these particles may benefit from the presence of EPS for two reasons. First, the exopolymers them-

selves can represent a source of highly labile particle-associated carbon. Second, since exopolymers often coat the surfaces of sediment particles, they bind and trap dissolved and colloidal compounds, making these compounds accessible (as particles) to particle-ingesting animals.

We conducted this study to address several parameters of exopolymers: whether capsule and slime EPS differed in digestibility to the deposit feeder Streblospio benedicti; whether more refractory types of EPS (e.g. capsules) could actually protect microbial cells from consumer digestion; and whether DOM adsorbed to EPS could be efficiently taken up by deposit feeders.

Materials and methods Culture and radiolabeling of bacterial cells

and exopolymers -Bacteria and exopolymers were extracted from cultures of the marine bacterium Pseudomonas atlantica (obtained in axenic culture from the American Type Cul- ture Collection, Washington, DC) and grown in seawater medium (25o/oo salinity) consisting of 0.5% (wt/vol) glucose and 0.25% Bacto-pep- tone. During early log phase of growth, either o-[U-14C]glucose [sp act 250-360 mCi mmol-‘; 3 mCi (100 ml)- ‘1 was added to label the cells and exopolymer fractions, or L-[U-14C]leucine [sp act > 300 mCi mmol- I; 500 &i (50 ml)- ‘1 (NEN) was added to label specifically the protein fraction of the EPS.

Isolation of slime exopolymers and bacterial cells-Capsule EPS are produced primarily during the log phase of growth and were harvested at that time. Slime EPS are produced mainly during stationary phase of growth and were harvested at that time. Slime EPS were isolated according to the methods of Kennedy and Sutherland (1987), using high-speed centrifugation (2 1,000 X g, 30 min) of stationary- phase cultures. Initially, Formalin (0.5% final concn) was added to preserve bacterial cells and to prevent loss of activity due to respiration of 14C0, by the cells during feeding experiments. The pelleted cells were washed three times in filter-sterilized (0.2~pm Nuclepore) seawater to remove residual Formalin. Feed- ing experiments indicated that the experimen- tal animals readily ingested such cells.

The supernatant containing the slime exopolymer was reduced to - 10% of its original

Exopolymer trophic interactions 1635

volume by freeze-drying. The exopolymer was then precipitated by extractions in cold (4°C) ethanol (70% final concn) for 8 h and alter- nately washed with ultraclean Milli-Q water for 1 h to remove small-molecular-weight ethanol-soluble components (Sutherland 1977). This.procedure was repeated three times. The exopolymer was then dialyzed exhaustively in Milli-Q water, freeze-dried, and stored (- 70°C) until use. The exopolymer was radioassayed in triplicate by dissolving a known quantity of dry 14C exopolymer in 100 ~1 of Milli-Q water and counting by liquid scintillation (see below). Activities were expressed as dpm 14C per bg dry exopolymer.

A portion of the 14C slime exopolymer was further extracted to reduce its protein content. Protein was extracted with the hot phenol method as described by Platt et al. (1985), al- though complete removal of protein is generally not possible without considerable loss of polysaccharide.

Isolation of capsule exopolymers and bacterial cells-Capsule EPS were isolated from cultures during log phase of growth. The presence of capsules was initially verified by light microscopy after wet-preparation negative staining with India ink (Duguid 195 1). Both the nature and thickness of capsules (or lack thereof) could be visually verified with light microscopy before each experiment.

Capsule EPS were harvested by initially fix- ing cell cultures in Formalin (0.5% final concn), then centrifuging (15,000 x g, 15 min) to remove any loose slime EPS. The supernatant was discarded and the pelleted cell fraction, containing the cells and their more tightly bound capsule EPS, was resuspended in filter- sterilized (0.2 .hrn) seawater. A small portion of these capsuled cells was used for feeding experiments (see below). We added 0.05 mM (final concn) ethylenediaminetetraacetic acid (EDTA) to the remaining cells. The cells were then centrifuged three more times (21,000 x g; 25 min), resuspending each time in the same supernatant. The multiple centrifugations provided enough shear force to remove most of the capsular material while leaving the bacterial cells intact. The capsule EPS, contained in the supernatant, were then isolated with the cold ethanol extractions and subsequent pu- rifications described above. Then the exopolymer was radioassayed.

Isolation of noncapsuled log-phase cells- The protective nature of the bacterial capsule against digestion was assessed by isolation of a noncapsuled version of the same bacterium for use in feeding experiments. The noncapsuled form was isolated by growing successive generations of the bacterium in medium com- paratively low in carbohydrate (i.e. Difco 22 16 media), then isolating the “C-cells” (cells that form crenatcd colonies on agar plates) (Bartlett et al. 1988). Cells stained with alcian blue (pH 2.5) and India ink had no visible capsule present. These cells were also labeled with [14C]glucose (NEN) and harvested during log phase of growth, similar to their capsuled counterparts described above.

Isolation of ‘“C-labeled DOM lysate- 14C- labeled DOM was derived from the marine- sediment diatom Amphora sp. (CCMP, Bige- low Lab. Ocean Sci.; Clone No. CCMPl389), isolated from Flax Pond sediments (near the Marine Sciences Research Facility, Stony Brook, New York). The diatom was cultured in f/2 medium with 25a/oo seawater (Guillard and Ryther 1962). During early log phase of growth, 1 mCi 50 ml-’ of NaHJ4C0, (sp act = 40-60 mCi mmol-l; NEN) was added to cultures. The cells were grown for about three generations to assure uniform labeling. r4C- labeled DOM lysate from the cells was then extracted. The cells were first fixed in Formalin (0.5% final concn), then centrifuged (5,000 x g; 15 min) twice, resuspending each time in seawater containing 0.05 M EDTA (final concn) for 1 h to remove any exopolymers and extracellular label associated with the cells. The cells were then suspended in seawater and lysed in a Polytron tissue homogenizer (Brinkmann Instr.). Cell fragments were separated from cy- tosolic lysate by centrifugation (30,000 x g; 30 min). The supernatant fraction was ultraf- iltered [30,000 nominal-molecular-weight limit (NMWL)] to remove any large-molecular- weight compounds. The filtrate was dialyzed (1,000 NMWL) in Milli-Q water. The isolated DOM (i.e. filtrate) was freeze-dried and stored (-70°C) until use. The activity of 14C-labeled DOM was measured in triplicate with liquid scintillation counting and expressed as dpm 14C per rug dry organic matter.

Adsorption of 5’Cr to exopolymers and cells- 51Cr was adsorbed to bacterial cells and exopolymers, and used in feeding experiments as


an inert tracer. Previous studies have shown that 51Cr(III) is adsorbed by deposit-feeding polychaetes with very low efficiencies (generally <3%, Lopez et al. in prep.). About 5 PC1 of 5’CrC13 in 0.5 M HCl was added (with an equal volume of 0.5 M NaOH) to a seawater suspension (0.5 ml) containing either 14C-labeled bacterial cells (Formalin-killed) or 14C- labeled exopolymer (dissolved in seawater, pH 7.8). The cells or exopolymer were then mixed with a small volume of sieved (~63 pm) natural sediment and inert fluorescent pigments. The 5 * Cr was allowed to adsorb for l- 1.5 h. The suspensions were then centrifuge-washed to remove nonadsorbed 51Cr and radioassayed.

Samples were first analyzed for 5 l Cr by gam- ma counting. Channel windows were set at 250- 400 keV. Samples were analyzed for 14C by liquid scintillation counting after adding 2 ml of ScintiVerse II (Fisher). Counting efficiency quench was estimated by the external stan- dards ratio method.

Composition of EPS -Purified samples, containing milligram quantities of exopolymer in a freeze-dried state, were weighed on a Cahn microbalance and then sealed in a helium atmosphere before combustion for C:H:N analysis with a Leeman Labs model CE 440 C:H:N analyzer. Precision was +0.3 wt% or better. Protein content of capsule and slime EPS was determined spectrophotometrically with the bicinchoninic acid protein assay (Pierce Chem. Co.) as described by Smith et al. ( 198 5). Bovine serum albumin was used as the standard.

14C-labeled DOM adsorption to exopolymer-Muddy sediment to be used in feeding experiments was collected from San Francisco Bay, cleaned twice for 8 h (25-30°C) in 30% hydrogen peroxide, and then in 0.5 M NaOH (w/Na,P,O, H20) to remove most organic matter. After each digestion the sediments were extensively washed in Milli-Q water to remove residual peroxide. About 0.25 g of the cleaned sediment was placed in seawater (25Ym) with dissolved exopolymer (10 pg exopolymer per g dry sediment) and incubated for 3 h (pH 7.9- 8.0; 22°C). Gentle agitation was used to facil- itate the formation of EPS coatings on the sediments. The sediments were then lightly centrifuged (500 x g; 3 min) to remove excess exopolymer in the supernatant and not bound to the sediment. The EPS-coated sediments

were mixed with 14C-labeled DOM (-0.135 &i) in seawater and incubated for 1 h. Pre- liminary experiments indicated that this incubation time was long enough to allow > 50% sorption of DOM to the sediment fraction. The sediments were once again lightly centrifuged to remove label not associated with the coated sediment. Fluorescent particles (Radiant Paint Co.) and 51 Cr (added as 1 &i 51CrC13) were also added as inert tracers and mixed with the sediments.

Cold-stage scanning electron microscopy of EPS-coated sediments-The presence of EPS coatings on the sediment particles was verified by observing samples of coated and uncoated sediment with scanning electron microscopy (SEM). Samples were prepared with the Sput- ter-Cryo technique (Robards and Crosby 1979). The specimens were observed with a cold-stage Cambridge stereoscan 250-mkz SEM at 15-kV accelerating voltage. Intact marine sediments, collected from an intertidal mudflat in San Francisco Bay near Palo Alto, were also observed for the presence of natural exopolymer biofilms. These sediments were collected from the sediment-water interface during low water. The biofilms on these natural sediments were qualitatively compared to exopolymer films artificially added to cleaned sediments.

Feeding on cells and EPS-The purpose of the feeding experiments was to determine whether different types of bacteria (capsuled and noncapsuled) or EPS (i.e. capsule and slime) were similarly digestible. The animal used in the feeding experiments was the spio- nid polychaete S. benedicti. It is a small, tu- bicolous deposit feeder commonly found in marine sediments near Flax Pond and many other locations on both east and west coasts of the U.S.

For feeding experiments, individuals (6-l 2 replicates per treatment) were placed in sep- arate microcosms containing the < 250~pm fraction of natural sediments. Therefore, grazing experiments were conducted in the presence of natural food resources. The animals were allowed to acclimate for at least 12 h to allow tube construction and to commence feeding on sediment. A small amount of 14C- labeled exopolymers, l 4 C-labeled bacterial cells, or exopolymer with adsorbed 14C-labeled DOM, previously mixed with sediment and inert fluorescent particles, was carefully added by pipette near the palps of a worm. The lo-


cation of the labeled food mixture could be simultaneously viewed under a dissecting mi- croscope as it was added. Once a portion of the labeled food mixture had been ingested, the remaining labeled sediment was carefully removed to stop ingestion of the labeled food mixture. This type of “pulsed addition” of labeled food has the advantage of allowing relatively constant feeding to occur on non- labeled food throughout the course of the experiment with only minor disturbance to the animals.

Because the fecal pellets of this polychaete are extruded on thC sediment surface outside the burrow, they could be easily observed and collected. Turnover times were measured by noting the first appearance of the fluorescently marked particles in the fecal pellets. Individual fecal pellets were collected by pipette until the fluorescent particles were no longer present in the newly produced fecal pellets and the gut was assumed to be empty of labeled food material. For S. benedicti adults, this generally took lo-40 min. Three unmarked pellets were also collected after the marked pellets to ensure that the gut was clear of labeled material. 14C and 51Cr radioactivity of food and individual fecal pellets was assessed as described above. At the end of an experiment, the polychaetes were fixed in Formalin (0.5% final concn), briefly rinsed in 0.1 mM HCl in seawater, and prepared for radioassay. The whole-body counts of these treatment worms were used later to assess the uptake of 14C-labeled DOM in solution (see below).

Uptake of solute DOM-The uptake of 14C- labeled DOM not directly associated with sediment or exopolymer (present in interstitial water) was controlled with a set of five live animals and five Formalin-killed animals. The animals were placed directly in Teflon wells containing seawater (pH 7.9-8.0) and 14C-labeled DOM (14,600 dpm ml-‘) but no sediments. The animals were incubated for 4 min (i.e. the maximum length ofthe hot feed period for any treatment worm). The added 14C-labeled DOM was collected from the labeled sediments used in treatment experiments. This 14C-labeled DOM represented that associated with the sediment that was in solution. The 14C-labeled DOM was collected by first resuspending the sediment in 1 ml of seawater, then lightly centrifuging the sediments and filtering (0.2 pm) the supernatant. After 4 min of in.-

cubation, the animals were removed, rinsed in seawater, and placed in unlabeled seawater for 30 min. The animals were fixed in Formalin (0.5% final concn) and briefly rinsed in 0.1 mM HCl in seawater. The whole-body radioactivity of these (control) polychaetes was compared with whole-body radioactivity from treatment polychaetes. Percent uptake due to DOM in solution was calculated as

% solute uptake whole-body 14C (from DOM solution exp)

= whole-body 14C (from DOM sediment exp)’

(1) Calculation of absorption eficiency- Per-

cent absorption efficiencies (%AE) were calculated according to Calow and Fletcher (1972) as modified by Lopez et al. (1989):

[ 14C : 51Crfccal] ?lbAE = l - [14C: 51Crlbod] x loo (2)

where 14C * 5 1 Cr . rood represents the ratio of 14C and 51 Cr labels in the food material and 14C : 5 1 Crrccal the ratio of these labels in the feces. Previous studies have shown no time-depen- dent absorption of either EPS or microbial cells (i.e. diatom cells) in S. benedicti (Lopez et al. in prep.).

Absorption efficiencies were normalized with arcsine transformations and analyzed with a one-way ANOVA for unequal sample sizes (Steele and Torrie 1980). A posteriori tests of significant terms were conducted with Dun- can’s multiple range tests (Steele and Torrie 1980). The mean gut passage times of polychaetes in experiments with 14C-labeled DOM adsorbed to EPS-coated treatment sediments vs. uncoated control sediments were compared with a t-test for unequal sample sizes (Steele and Torrie 1980). Uptake of 14C-labeled solute by live polychaetes and Formalin-killed polychaetes was compared with a t-test for equal sample sizes (Steele and Torrie 1980). All data transformations and analyses were conducted with the SYSTAT statistical analysis system (Wilkinson 1987).

Results Isolation of EPS and 14C-labeled DOM-

The yield of slime EPS from stationary-phase cultures ofP. atlantica was 2.7OkO.66 mg ml-’ culture medium (Table 1). The activity of the purified exopolymer was 4,336 + 42 1 dpm


Table 1. Compositions, yields, and ‘VI activities of isolated exopolymers and “Cr adsorbed to exopolymers. Yields (n = 3j, activity, and C: N (n = 2) values represent the mean i SE. Yield values are given as mg dry EPS mlP’ culture media. Activities are given as dpm mg 1 dry EPS; nd--not determined.

2.01+m 0.68 2.6

0.36+0.1 0.32 39.2

nd

8.94

nd

4.3*0.1

Capsule EPS (w/specific protein labeled) 0.27kO.2 0.88 nd nd nd

rng-l dry EPS. About 58% of the ‘Y label recovered from cultures was associated with the slime EPS fraction. The C : N ratio of purified slime EPS was 17.6kO.6. The mean protein content was 9.3 wt%. The activity of the same exopolymer after removal of protein de- creasedto3,998?375 dpmmg-‘dryEPS. The remaining mean protein content of this polymer was 2.6% (Table 1). Cells collected from slime-producing cultures did not have observ- able capsules.

The yield ofcapsule EPS from log-phase cultures of P. atlantica was 0.36iO.12 mg ml ~’ culture medium (Table 1). Capsule EPS SW- rounding log-phase cells were readily visible with light microscopy after staining with India

ink (Fig. 1). Activities of exopolymer derived from [14C]glucose-labeled cultures was 2,644i486 (n = 3) dpm mg-’ dry exopolymer. About 31% of the recovered 14C was incor- porated in capsule exopolymer material. The C: N ratio of the purified capsule EPS was 4.27t0.03, and mean protein content was 39.2 wt%. The activity of capsule exopolymer isolated from cultures containing [‘4C]leucine (to label the protein fraction of the exopolymer) was 872i 13 1 dpm mgm’ dry EPS. About 24% of the ‘Y recovered was associated with exopolymer material; the rest was associated with cells. Activities of L4C-labeled bacterial cells are given in Table 2. The yield of 14C-labeled DOM cellular lysate was 11 &g of dry organic

b’

Fig. 1. Micrographs showing different cell types of the bacterium &wdomonas atlantica (stained with India ink) used in feeding experiments with the polychaete Streblospio benedicli: A-bacterial cell with an enopolymer capsule from log phase of growth; B-bacterial cell during stationary phase of growth which produces slime EPS (not visible); C-noncapsuled (C-cells) cell from log phase (see text for explanation). (b-Bacterial cell: e-enopolymcr capsule.) Scale bars = I ,,m.

E’xopolymer trophic interactions 1639

Table 2. Activities of 14C- and 5’Cr-labeled bacterial cells of Pseudomonas atlantica (with adsorbed 5’Cr) used in feeding experiments with the polychaete Streblospio benedicti. Activity values are given as the mean k SE (n = 3).

Table 4. Feeding by the polychaete Streblospio benedicti on various “C-labelcd bacterial cells (Pseudomonas atlantica). All values given as mean + SE. Mean values for percent absorption efhciency with the same letter superscript arc not significantly different (P > 0.05).

Bacterial ccl1 types “‘C activily (dpm cell ‘) Bacterial cell types

Gut-passage No. of time (min) animals % AE

Stationary-phase cells (noncapsuled)

Log-phase cells (capsuled)

Log-phase cells (noncapsuled)

36+11 0.53

58+24 0.76

27a 17 0.34

Capsuled cells (log phase) 19k7.8 9 5.1 +2.1a

Noncapsuled cells (stationary phase) 17k8.3 6 17.3k7.1”

Noncapsuled cells (log phase) 17Ik9.3 7 14.4t-2.1”

matter from 50 ml of culture. The activity of the 14C-labeled DOM was 59,091 dpm r.cg-l

Feeding on EPS and cells -Adsorption of dry organic matter.

51Cr to EPS used in feeding experiments re- sulted in 14C : 51Cr ratios between 0.32 and 0.88 (Table 1). The mean percent absorption efficiencies (%AE) of S. benedicti individuals fed different exopolymer treatments are given in Table 3. ANOVA with arcsine-transformed data showed a significantly (P < 0.001) lower absorption of capsule EPS than of slime EPS.

The activities and 14C : 51Cr ratios of bacterial cells used in feeding experiments are given in Table 2. Gut-passage times of S. benedicti individuals ranged from 6 to 49 min during feeding experiments and were not significantly different (P > 0.05) among any of the bacterial cell types on which the animals were fed. The %AE of S. benedicti fed different bacterial cell types ranged from 5 to 17% (Table 4). There was significantly (P < 0.00 1) lower absorption of capsuled cells from log-phase cultures than noncapsuled cells from either log- or station-

ary-phase cultures. No difference was detect- able between noncapsuled cells from log-phase

Cold-stage SEM of EPS-coated sediments - Examinations of EPS-coated sediments (to be

and stationary-phase cultures (P = 0.54).

used in adsorbed-DOM feeding experiments) with cold-stage SEM showed conspicuous and abundant coatings of EPS closely surrounding the individual sediment grains (Fig. 2). The polymer fibrils formed a complex intermeshing network. Cleaned control sediments had no coatings. Examination of natural sediments with this technique (Fig. 3) showed that exo- polymerlike biofilms were abundant in sediments at the sediment-water interface, where surface deposit feeders ingest sediment. These natural biofilm coatings were qualitatively similar to the biofilms artificially produced on sediments with microbial EPS.

Feeding on DOM adsorbed to EPS- 14C-labeled DOM in solution, extracted from treated sediments, was 14,600 dpm ml-‘. This con- centration represented 7.7% of the 14C-labeled DOM initially adsorbed to the sediments. 14C-

Table 3. Feeding by the polychaete Streblospio benedicti on capsule vs. slime EPS. All values are mean + SE. Mean values for percent absorption efficiency with the same letter superscript are not significantly different (P .> 0.05).

Exopolymcr Gut-passage time

(min) No. of

animals “h AE

Slime EPS “intact” (uniformly labeled)

Slime EPS (polysaccharide extract)

Capsule EPS “intact” (uniformly labeled)

Capsule EPS (protein labeled)

22+ 10.2 11 72f 1 la

19k5.9 8 86f8”

23k8.8 8 48+11b

26k 10.0 7 21k15c


Fig. 2. SEM micrographs of uncoated control sediments (above) and EPS-coated (treatment) sediments (below). These types of sediments were used to bind ‘Y-labeled DOM for the feeding experiments. The sediments were initially cleaned to remove most organic material (see methods). Note abundant exopolymer biofdms on coated sediments. SEM was conducted with Sputter-Cryo technique (see methods) to allow exopolymcn to remain in a hydrated state during observation. (S--Sediment grain; EPS-enopolymcr biofilm.) Scale bar = 10 WI.

labeled DOM uptake from solution was not significantly different (P = 0.867) between live (n = 5) and Formalin-killed (n = 5) polychaetes over the incubation period. These two groups of animals were therefore pooled during later ANOVA analyses. Results of ANOVA with arcsine-transformed data of %AE showed a highly significant treatment effect (P i 0.001). A posteriori analyses indicated that the mean %AE of DOM on EPS-coated sediments (%AE = 79%) was significantly higher than on uncoated control sediments (%AE = 32%) (Table

5). Mean uptake due to solute DOM (%AE = 4%) was significantly less than uptake from both types of sediments. Gut-passage times were not significantly different (P = 0.752) between EPS-coated treatment sediments (X = 18.5k2.8 min, n = 10) and uncoated control sediments (.a? = 2O.Ok3.8 min, n = 7).

Discussion Exopolymers have been previously recog-

nized as a highly labile food source for animals. The present study, however, shows that not all

Empolymer trophic interactions 1641

Fig. 3. SEM Francisco Note the presence of abundant polymeric biohlms coating the sediment panicles which obscure the view afbac and diatom cells embedded within them. (D-Diatom cell: EPS-enopolymerlike biohlm.) Scale bar = 10 @m.

Bay. mid

of them are easily digestible and suggests several trophic roles for these secretions. They can range from a labile food source to a refractory barrier against the digestion of microbial cells, depending on their composition. In addition, the exopolymer matrix represents an active adsorptive microenvironment for the concentra- tion of DOM from the surrounding water. Such sorption processes make this DOM highly accessible to consumer animals ingesting the sediment.

In P. atlantica, capsule EPS isolated from

log-phase cells and slime EPS isolated from stationary-phase cells had very different protein contents and C: N ratios (Table 1). The tight capsule EPS had a high protein content, whereas the loose slime EPS were largely composed of polysaccharides (Table 1). Subse- quent feeding experiments with the S. benedicti showed that loose slime EPS were used with very high absorption efficiencies, whereas capsule EPS were used relatively poorly. The enhanced utilization of slime EPS appeared to reflect their overall high content of polysac-


charides and correspondingly lower protein content compared to capsule EPS. Experi- ments in which the protein portion of the EPS was removed or specifically labeled supported these results further. Additional experiments indicated that bacterial cells with exopolymer capsules were digested less efficiently than cells with no capsules or very reduced capsules. These data support the idea that capsule EPS may protect microbial cells against digestion by grazers.

Many slime EPS are secreted in copious amounts resulting from what is termed a “met- abolic excess response” or an “unbalanced growth response” (Sutherland 1977). This response occurs when there is an abundance of one nutrient, usually carbon, while other nutrients necessary for balanced cell growth are limiting. The excess carbon is metabolically shunted by the cell into the formation of extracellular polymers which are typically high in polysaccharide but low (or lacking) in protein. Such conditions are thought to occur pe- riodically in nature (Costerton et al. 1985).

The second type of polymer, a capsule, is secreted for very different reasons. In natural sediments, microbial cells secrete capsules as a protective response to various stressful conditions and to stabilize the pH and osmotic environment in proximity to the microbial cell (Costerton 1974). Within the guts of vertebrate animals, resident microbial flora have exopolymer capsules that are specifically designed to resist digestive enzymes and overall harsh conditions. This process has been closely examined in the study of rumen bacteria in cattle (Costerton et al. 1985). Rumen bacteria create colonies enclosed in biofilms and capsules on the surfaces of cellulose particles. The capsules provide protective effects for the cells, and the biofilms help the microbial cells concentrate their enzymatic digestion of the cellulose. Once the cellulose particle has been degraded, the whole consortia eventually disintegrates and is itself digested.

Capsule EPS have varying amounts of protein (Sutherland 1977). Proteins are present in EPS in several forms, ranging from complex structural glycoproteins to more loosely associated exoenzymes and peptides. The form of the protein, however, may be especially important. For example, glycoproteins may represent key components in determining the

structural integrity and protective nature of the bacterial capsule. Glycoproteins maintain the tight tertiary structure of vertebrate intestinal mucilage (Silberberg 1989) which is needed to constantly resist enzymatic digestion by hy- drolases (Sellers et al. 1988). In a similar man- ner, the tighter tertiary structure of bacterial capsules may slow (or prevent) digestive enzymes from reaching the cells during gut passage (sensu Bacon 1979). Under most conditions in marine sediments, however, the bacterial capsule probably represents an ad- aptation designed to protect cells from adverse microenvironmental conditions other than grazing.

In the present study there were two reasons to compare feeding of capsuled (log phase) and noncapsuled cells (stationary phase): to provide a direct measure of the digestibility of bacterial cells in these different physiological states; and to allow the protective nature of the capsule EPS to be examined while they were present in their intact form surrounding the cell. These observations were important because the exopolymer extraction processes used to purify the polymer (i.e. centrifugation, ethanol precipitations, etc.) might modify the tertiary structure of the polymer, potentially affecting its digestibility. The use of cells with intact capsules demonstrated that the protective effect and refractory nature of the capsule was not significantly affected by the extraction process. The insignificant difference in digestibility between log-phase cells without capsules and stationary-phase cells with slime indicated that in these experiments differences in the physiological state of the cells did not significantly influence their digestibility.

Capsule thickness and physical integrity can vary considerably. Capsules can range from only lo2 nm (distinguishable only by electron microscopy) to > 10 pm (Sutherland 1977). If such reduced capsules were present on control cells, they appeared to provide no protective effect from digestion processes, as evidenced by their similar digestibility to stationary-phase cells.

Natural sediments were examined by cold- stage SEM. This technique preserves the intact hydrated structure of exopolymers which is easily destroyed. These observations showed that biofilm matrices commonly coat intertidal sediments where deposit feeders reside (Fig.


3). Additions of purified EPS to precleaned sediments formed coatings that were qualitatively similar to the natural biofilms (Fig. 2). These sediments represented a useful substra- tum for examining exopolymer-feeding interactions. In addition to providing an abundant source of potentially labile carbon, the polymer fibrils form a complex intermeshing network possessing abundant sites for adsorption of DOM in natural sediments.

Adsorption of DOM to sediments is a common process. A large portion of the total organic matter in sediments is thought to be present mainly as adsorbed material on mineral surfaces (Mayer et al. 1985). DOM readily binds to different components of marine sediments and detrital particles. The present study demonstrates that DOM associated with the mineralogical portions of sediments can be used efficiently by deposit feeders. The conversion of DOM in solution to a particulate form (via binding to sediments) increased its availability to the polychaete.

S. benedicti used DOM associated with two different components of sediment particles. DOM bound directly to peroxide-cleaned sediment (Fig. 1A) was used, but with a relatively low absorption efficiency (%AE = 32%; Table 2). There are many different organic and inorganic components of sediments that can significantly bind DOM. The mineralogical portion of the sediment itself has many different components (e.g. hydrous oxides, silicates, car- bonates) (Luoma and Davis 1983) that have different capacities for binding DOM. In the present study, it is postulated that the DOM bound directly to mineralogical portions of sediments (i.e. cleaned sediment) was probably less available because it was more tightly bound and could not be removed easily from the sediment during gut passage.

The additional presence of an exopolymer biofilm on the sediment (Fig. 1B) significantly increased the availability of the bound DOM to the polychaete. When DOM was bound to EPS surrounding the sediment, absorption efficiencies of the DOM more than doubled. The higher absorption efficiencies probably oc- curred because the exopolymers are themselves highly labile, and organic coatings (such as exopolymer biofilms) on the sediments reduce the binding strengths of DOM to the sediment. These properties of EPS probably in-

creased the availability of the bound DOM to animals ingesting the sediment. Such binding processes can make DOM potentially available to particle-ingesting animals, provided the DOM can be removed from the particles during gut passage.

About 4% of the 14C-labeled DOM uptake by S. benedicti in sediment-feeding experiments may have been due to solute uptake (i.e. not associated with sediments). These values probably overestimate the actual uptake of solute DOM because during solute control experiments the whole body of the animal was exposed to 14C-labeled DOM, whereas during actual feeding experiments only the palp and head area of the animal were exposed to potential solute label. Overall uptake due to DOM in solution, therefore, appeared to be a very minor avenue for organic matter uptake (Table 2).

DOM can be transferred to consumer animals by three major pathways. The first involves its direct uptake by animals from solution. The ability of many animals to efficiently take up DOM from solution, however, appears limited (Gomme 1982). A second, indirect pathway involves the microbial mineralization of DOM followed by ingestion of the microbial cells by consumers. This pathway is also inefficient because during microbial metabolism substantial organic matter (i.e. 50%) is lost due to excretion and respiration (Ducklow et al. 1986). A third pathway involves the recent adsorption of DOM to particles and its consumption by animals. This pathway is potentially efficient because DOM is transferred directly to animals, with no associated losses via microbial metabolism, and uptake can be highly efficient because the DOM is associated with particulate forms.

The exopolymer matrix, when present on particle surfaces, represents an efficient microenvironment to directly transfer DOM to animals for several reasons. First, DOM readily associates with particulate EPS either through binding or colloidal trapping. Second, the exopolymers can themselves represent highly labile carbon sources for animals, and they are coincidently ingested with other sediment-associated components during normal feeding. Also, even if the exopolymer itself is not labile, the DOM adsorbed to these polymers might still be available to animals. This availability


occurs because much of the binding is a result of the abundant carboxyl ligands on acidic polysaccharides of the exopolymer. These ligands can be relatively weak (Sutherland 19 8 9), so adsorbed compounds can be easily disso- ciated from the polymer during gut passage. It should be emphasized that the DOM (i.e. diatom lysate) used in our experiments is highly labile. The specific binding capacities of other types of DOM with sediments will depend on the type of DOM and its complex interactions with various organic and inorganic sediment components, all of which can potentially influence DOM bioavailability (Luoma and Da- vis 1983).

In water-column systems, exopolymer- bound DOM should be transferred with equal facility to animals consuming water-column aggregates, since the amorphous detritus of marine snow aggregates is often composed of microbial EPS. As cells within an aggregate are lysed, the surrounding exopolymer matrix will bind the DOM being released, slowing its dif- fusive loss. This will keep nutrients within the aggregate environment where they can enhance recycling or serve as direct food sources for the many protozoans and meiofaunal animals observed in aggregates (Shanks and Ed- mondson 1990). One important trophic role of EPS, therefore, may be in influencing the accessibility of DOM to animals, especially those which key on particulate foods.

Many bacterial cells secrete capsule EPS during periods of active growth. Such a mechanism has implications in the use of microbial cells as food by consumers. Once in the gut of a deposit feeder, actively growing capsuled cells may better survive digestion than noncapsuled senescent cells. The vast majority of bacterial cells in marine sediments are thought to exist in a relatively inactive state (Novitsky 1983; Carman 1990; and others). Therefore, such mechanisms can influence the utilization of microbial cells in lower food webs and the passage of carbon and nitrogen to higher trophic levels, especially in heavily grazed areas.

A final implication of this work concerns symbiotic and pathogenic bacteria. Many bacteria use capsule EPS to evade phagocytosis by host defenses (see Dudman 1977). As these symbiotic cells reach less active physiological states, the capsule integrity which protects them will degrade and the cells will be more readily

recognized and phagocytized by the host. Such a mechanism would provide a means by which populations of symbiotic microorganisms could be regulated by the host. Small populations of symbionts which underutilize available nutrients could be allowed to proliferate as actively growing capsuled cells. In contrast, very large populations of cells that have be- come nutrient limited and senescent will be reduced in numbers until an actively growing population near equilibrium is reached.

Our data predict that the digestibility of EPS as a food source is influenced, in part, by their protein content. Polymers that are compara- tively rich in protein (perhaps glycoproteins) protect microbial cells because they are more refractory to digestion. In contrast, many slime EPS, which are components of biofilms, may represent a highly labile carbon source for animals because they are more abundant in easily hydrolyzable polysaccharides. The role of EPS as an N source will probably depend on N compounds that are adsorbed to or more loosely associated with the exopolymer (e.g. exoenzymes) rather than those that are part of its structure. The roles of other interacting factors in influencing the physical integrity and refractory nature of exopolymers, however, must also be considered.

References BACON, J. S. D. 1979. Factors limiting the action of

polysaccharide degrading enzymes, p. 269-284. Zn R. C. W. Berkeley et al. [eds.], Microbial polysaccharides and polysaccharases. Academic.

BAIRD, B. H., AND D. THISTLE. 1986. Uptake of bacterial extracellular polysaccharides by a deposit-feeding ho- lothurian (Isostichopus badionotus). Mar. Biol. 92: 183- 187.

BARTLETT, D. H., M. E. WRIGHT, AND M. SILVERMAN. 1988. Variable expression of extracellular polysaccharide in the marine bacterium Pseudomonas atlantica is controlled by genome rearrangement. Proc. Natl. Acad. Sci. 85: 3923-3927.

CALOW, P., AND C. FLETCHER. 1972. A new radiotracer technique involving 14C and 5’Cr for estimating as- similation efficiencies of aquatic primary consumers. Oecologia 9: 155-170.

CARMAN, K. R. 1990. Radioactive labeling of a natural assemblage of marine sedimentary bacteria and mi- croalgac for trophic studies: An autoradiographic study. Microb. Ecol. 19: 279-290.

CHRISTENSEN, B. E., J. KJOSBAKKEN, AND 0. SMIDSROD. 1985. Partial chemical and physical characterization of two extracellular polysaccharides produced by marine periphytic Pseudomonas sp. strain NCMB 202 1. Appl. Environ. Microbial. 50: 837-845.


COSTERTON, J. W. 1974. Structure and function of the cell envelope ofgram-negative bacteria. Bactcriol. Rev. 38: 87-l 10.

-, T. J. MARRIE, AND K. J. CHENG. 1985. Phcnom- enon of bacterial adhesion, p. 3-43. In D. C. Savage and M. Fletcher [eds.], Bacterial adhesion. Plenum.

DECHO, A. W. 1990. Microbial cxopolymcr secretions in ocean environments: Their role(s) in food webs and marine processes. Oceanogr. Mar. Biol. Annu. Rev. 28: 73-l 53.

- AND R. CASTENHOLZ. 1986. Spatial patterns and feeding of meiobenthic harpacticoid copepods in rc- lation to resident microbial flora. Hydrobiologia 131: 87-96,

- AND D. J. W. MORIARTY. 1990. Investigation of bacterial mucus-exopolymer utilization by marine animals. Methodology and results using harpacticoid copepods. Limnol. Oceanogr. 35: 1039-l 049.

DUCKLOW, H. W., D. A. PURDIE, P. J. LEB. WILLIAMS, AND J. M. DAVIES. 1986. Bacterioplankton: A sink for carbon in a coastal marinc plankton community. Sci- ence 232: 865-867.

DUDMAN, W. F. 1977. The role of surface polysaccharides in natural environments, p. 357-414. In I. W. Sutherland [ed.], Surface carbohydrates of the pro- karyote cell. Academic.

DUGUID, J. P. 195 1, The demonstration of bacterial capsules and slime. J. Pathol. Bacterial. 63: 673-685.

EPP, R. W., AND W. M. LEWIS. 198 1, Photosynthesis in copepods. Science 214: 1349-l 350.

GOMME, J. 1982. Epidermal nutrient absorption in marine invertebrates: A comparative analysis. Am. Zool. 22: 69 l-708.

GUILLARD, R. R. L., AND J. H. RYTHER. 1962. Studies on marine planktonic diatoms. 1. Cyclotella nana Hustedt and Detonula confervacea (Cleve) Gran. Can. J. Microbial. 8: 229-239.

KENNEDY, A. F. D., AND I. W. SUTHERLAND. 1987. Anal- ysis of bacterial exopolysaccharides. Biotcchnol. Appl. Biochem. 9: 12-19.

LOPEZ, G. R., P. TANTICHODOK, AND I.-J. CHENG. 1989. Radiotracer methods for determining utilization of sedimentary organic matter by deposit feeders, p. 149- 170. In Ecology of marinc deposit feeders. Coastal Estuarine Stud. V. 31. Springer.

LUOMA, S. N., AND J. A. DAVIS. 1983. Requirements for modeling trace metal partitioning in oxidized estuarine sediments. Mar. Chem. 12: 159-181.

MAYER, L. M., AND OTHERS. 1985. Biological and gran- ulometric controls on sedimentary organic matter of an intertidal mudflat. Estuarine Coastal Shelf Sci. 20: 49 l-503.

NOVITSKY, J. A. 1983. Heterotrophic activity throughout a vertical profile of seawater and sediment in Halifax Harbor, Canada. Appl. Environ. Microbial. 53: 2368- 2372.

PLATT, R. M., G. G. GEESEY, J. D. DAVIS, AND D. C. WHITE. 1985. Isolation and partial chemical analysis of firmly bound exopolysaccharidc from adherent cells of a freshwater sediment bacterium. Can. J. Mi- crobiol. 31: 675-680.

PORTER, K. G. 1976. Enhancement of algal growth and productivity by grazing zooplankton. Science 192: 1332-1334.

REES, D. A., E. R. MORRIS, D. THOM, AND J. K. MADDEN. 1982. Shapes and interactions of carbohydrate chains, p. 196-29 1. Zn G. 0. Aspinall [ed.], The polysaccharides. V. 1. Academic.

ROBARDS, A. W., AND P. CROSBY. 1979. A comprehen- sive freezing, fracturing and coating system for low- temperature scanning electron microscopy. Scanning Electron Microsc. 2: 325-344.

SELLERS, L. A., A. ALLEN, E. R. MORRIS, AND S. B. ROSS-MURPHY. 1988. Mucus glycoprotcin gels. Role of glycoprotein polymeric structure and carbohydrate side chains in gel formation. Carbohydrate Res. 178: 93-l 10.

SHANKS, A. L., AND E. W. EDMONDSON. 1990. The vertical flux of metazoans (holoplankton, meiofauna, and larval invertebrates) due to their association with marine snow. Limnol. Oceanogr. 35: 455-463.

SILBERBERG, A. 1989. Mucus glycoprotein, its biophys- ical and gel-forming propcrtics, p. 43-63. Zn Mucus and related topics. Symp. Sot. Exp. Biol. No. 43. Res. Books.

SMITH, P. K., AND OTHERS. 1985. Mcasuremcnt of protein using bicinchoninic acid. Anal. Biochcm. 150: 76-85.

STEELE, R. G. D., AND J. H. TORRIE. 1980. Principles and procedures of statistics, 2nd ed. McGraw-Hill.

SUTHERLAND, I. W. 1977. Bacterial exopolysaccharides-their nature and production, p. 27-96. In I. W. Sutherland [ed.], Surface carbohydrates of the pro- karyote cell. Academic.

-. 1989. Microbial polysaccharides-a comparison with cukaryote polymers, p. 389-402. In Mucus and related topics. Symp. Sot. Exp. Biol. No. 43. Res. Books.

WILKINSON, L. 1987. SYSTAT: The system for statistics. Systat Inc.

Submitted: 18 March 1992 Accepted: 1.5 March 1993

Revised: 6 April 1993

Exopolymer microenvironments of microbial flora: Multiple and interactive effects on trophic relationships

Documents