sign in sign up
<<
Home , Zebra mussel

AQUATIC Vol. 21: 187-194.2000 1 Published March 31 Aquat Microb Ecol

Effects of the zebra on nitrogen dynamics and the microbial community at the sediment-water interface

Peter J. Lavrentyevll*, Wayne S. ~ardner~l*, Longyuan Yang3

'University of Michigan. Cooperative Institute for and Ecosystem Research. 2200 Bonisteel Boulevard. Ann Arbor. Michigan 48109. USA 2National Oceanographic Atmospheric Administration, Environmental Research Laboratory, 2205 Commonwealth Boulevard, Ann Arbor, Michigan 48105, USA 'Chinese Academy of Sciences, Nanjing Institute of Geography and Limnology, 73 East Beijing Road, Nanjing, Jiangsu 21008, PR China

ABSTRACT: A flow-through expenment was conducted on intact cores of sedunents from Saginaw Bay, , to examine how trophic interactions between filter-feeding bivalve and microbial populations could affect nitrogen dynamics at the sediment-water interface. The zebra mus- sels used in this experiment removed a large proportion of protozoa and from the over- lying water, particularly heterotrophic nanoplankton (up to 82 %), while bacterial population~showed less change. A 3-fold decrease in the protozoan to bacterial carbon ratio corresponded to a 2.5-fold increase in relative ammonium removal rates as estimated from the dark loss of '5N-ammonium. Excre- tion by the bivalves also increased net ammonium flux to the water, thus elevating the total calculated area1 ammonium removal rates to about 6-fold over rates observed in the control treatment. These data suggest that filter-feedng bivalves may significantly affect nitrogen transformation rates near the sed- iment-water interface by excreting ammonium and altering the microbial structure at the sediment-water interface.

KEY WORDS: Nitrogen . . Sediment-water interface . Bivalve mussels

INTRODUCTION If oxygen is present, as often is the case in shallow waters, a portion of ammonium may be oxidized to The sediment-water interface is a region of active nitrate (Ward 1986). In turn, the nitrate may be com- biological and biogeochernical processes, such as labile pletely or partially denitrified to nitrogen gas in adja- organic matter biodegradation and ammonium pro- cent sites (Seitzinger 1988). Understanding the mecha- duction in shallow coastal systems. Under such condi- nisms that control the extent of N transformations at tions, kinetic diffusion-reaction models based on equi- the sediment-water interface is important, because libration of ammonium between the sediments and these processes can affect the amount of mineralized N pore water do not adequately predict actual ammo- that is ultimately available to primary producers or lost nium flux into overlying waters (Ullman & Aller 1989). through denitrification. Trophic cascades involving bacterivorous protozoa can affect community N regeneration rates (Suzuki et Present addresses: al. 1996) and bacterial community structure in the water 'University of Akron, Department of Biology, Akron, Ohio 44325-3908, USA. E-mail: [email protected] column (Jiirgens & Giide 1994). Furthermore, plank- "University of at Austin, Marine Science Institute, 750 tonic bacterivorous protozoa can inhibit the biogeo- Channel View Drive, Port Aransas, Texas 78373, USA chemical process of nitrification, directly by reducing

0 Inter-Research 2000 188 Aquat Microb Ecol21: 187-194,2000

bacterial numbers and indirectly by forcing bacteria to MATERIAL AND METHODS grow in colonies, whereas predation by larger zoo- diminishes these effects (Lavrentyev et al. Material for this study was collected from Saginaw 1997). The important role of the bottom-dwelling pro- Bay, a large (about 82 km long and 42 km wide) bay tozoa in consuming heterotrophic bacterial production extending off the western edge of Lake Huron. The (e.g. Bak et al. 1995) supports the idea that similar cas- shallow (1 to 3.5 m depth) inner bay receives large cading trophic effects on nitrogen transformation may inflows of enriched waters from the Saginaw River. take place at the sediment-water interface. Here, rnicro- Since 1991, the zebra mussel has heavily infested the organisms are involved in complex interactions with inner bay (Nalepa et al. 1995). other benthic organisms that have significant effects Intact cores of sediments and the overlying water on nitrogen transformations (Dame et al. 1984, 1991, were collected by SCUBA divers from a site near Au Henriksen & Kemp 1988). Gres Point (43'56'N, 83 "52'W) on June 30, 1995. One of these organisms is the zebra mussel Dreis- These samples were taken using 76 mm internal diam- sena polymorpha, an invasive filter-feeding bivalve that eter plastic tubes. Simultaneously, several pebbles col- alters the structure of food webs in shallow areas by onized by the zebra mussel were collected for experi- fiitering particies (e.g. Naiepa et al. 1995) and excret- mentai additions. Severai 20 i carboys or surface water ing nutrients (Gardner et al. 1995b. Binelli et al. 1997). from the same site were collected to supply a flow- In this study, we examined how trophic interactions, through system. The samples were transported to the induced by zebra mussels, can modify the microbial laboratory in insulated coolers within a few hours after food web structure and nitrogen transformations at the collection. sediment-water interface. The sedirnents were placed in a system consisting Studying trophic interactions and nutrient transfor- of the cores (in the original collection tubes), a clear mations involving microorganisms at the sediment- polycarbonate vessel with intake lake water, a Techni- water interface presents methodological difficulties. con peristaltic pump, and an outflow collection vessel The use of a I5N tracer is necessary to study N dynam- (Fig. 1). This laminar flow system is an improved and ics because of close coupling between ammonium re- larger version of the flow-cell described earlier (Gard- generation, nitrification, and denitrification. The mass ner et al. 1991). It was designed with an 0-ring seal and analysis techniques for 15N require large sample vol- plunger that could be lowered to the desired position umes and involve multi-step sample preparations that without disturbing the surface sedirnents. The overflow increase the risk of contamination. The physical, che- tube released water as the plunger was positioned but mical and biological nature of the sedimentary matrix was closed during operation. Water was pumped over is impossible to replicate in the laboratory; yet, the the sediment surface at a rate of 1 m1 rnin-l. maintenance of this structure is crucial to many ben- The pebbles holding the zebra mussels were of dif- thic processes (Miller-Way & Twilley 1996).The neces- ferent size and shape and had large amounts of peri- sity to maintain sediment integrity during an experi- phyton on them. Mussels were removed from the peb- mental incubation huts the application of conventional bles by cutting the byssal threads with a razor blade size-fractionation and dilution techniques that are com- monly used to modify food web structure in planktonic Hand'e Overflow tube Inflow - studies. Further, the presence of meroplanktonic pro- 0 tozoa, which are able to migrate freely between sedi- ments and the , may greatly diminish the Outflow effects of experimental manipulations with inlet water in such studies. +- 0-ring The purpose of this study was to examine how trophic + Lake I interactions between filter-feeding bivalve mussels and water microbial populations could change nitrogen dynamics in controlled experimental systems using intact sedi- ments. To simulate a trophic cascade, we used the ri zebra mussel as a 'biological filter' in a sedi- ment core continuous flow cell. We used this approach in combination with high performance liquid chromato- graphic analysis of ammonium concentration and atom U Intake water Collection vessel % 15N to test the hypothesis that cascading trophic Fig. 1. The continuous flow through system. The diameter of effects involving bacterivorous protists affect nitrogen the core was 76 mm. The flow rate for the unfiltered water transformation at the sediment-water interface. was 1 m1 mix' Lavrentyev et al.: Casce ~dlngeffects on N dynamics 189

and rinsed with bay water to remove periphyton before passage across the dark core should not have ex- they were used in the experiment. Two adult mussels, ceeded that occurring in the inlet water held under shell length 22.3 mm * 0.69 SE, were added to each light/dark conditions. Therefore, the observed differ- triplicate core. By the time the system had reached ences in 15NH4+concentrations between the inflow and steady state, the zebra mussels remained either on the outflow waters were assumed to result from microbial top of the sedlments or attached to the side of the processes at or near the sediment-water interface. core tube above the sediment surface. Thus, they were Results from the flow-through core system, without positioned to filter overlying water rather than pore (incubation Days 2, 3, and 4) and with (incubation water. All zebra mussels remained alive during the Days 5, 6, and 7) I5NH4+additions, were used to cal- experiment and were still filtering water at the end of culate ammonium removal rates. Successive results the incubation. from each core were averaged before (n = 3) and after The cores and the inflow water from the sample site (n = 3) addition of I5NH4+to yield core-specific ammo- were incubated in an indoor Percival incubator that nium concentration and isotope ratio values. The core- was set for a 12:12 h light:dark cycle and held at specific values were than averaged among replicate ambient temperature (21°C). The inlet water was held core treatments to calculate mean values and standard under fluorescent Lights to ensure phytoplankton errors (SE) for the zebra mussel and control treatments, growth. Triplicate cores with mussels and triplicate respectively, before and after 15NH4+ additions. We controls without mussels were each assembled and assumed that the same percentage of ammonium was wrapped in aluminum foil to prevent light penetration removed, whether it was released from the sediments, and make it possible to differentiate between removal excreted by organisms, or added to the inlet water as by nitrificatiorddenitrification and algal uptake. Large 15NH4+.Net areal ammonium flux (N,,,, pm01 N m-2 h-') quantities of light (both visible and UV lengths) can was calculated as follows: inhibit nitrification via photooxidation of bacterial N,,, = (No N,) X flow rate/core surface area cytochrome c (Bock et al. 1989). - = (No- NI) X 13.3 The concentrations of NH4+ and nitrates (equals (1) NO2-+ NO3-) were monitored over time. Ammonium where No and NI were ammonium concentrations (PM) concentrations and 15NH4+ isotope ratios were res- in the outflow and inflow waters, respectively, during pectively measured via 2 high performance liquid chro- Days 2,3 and 4, before the addition of 15NH4+.The flow matography methods (Gardner & St. John 1991, Gardner rate was 0.060 1 h-' and the surface area of the core et al. 1995a). Concentration of nitrates was measured us- was 0.0181 m2. Total ammonium flux (N,,,,,, pm01 N ing a standard colorimetric technique (Strickland & Par- m-2 h-'), i.e. N,,, corrected for estimated removal, was sons 1972) with a Technicon autoanalyzer. then calculated as: After the system had been maintained at steady state (based on the concentrations of dissolved nutnents) for 3 d, 15N-ammonium isotope (10 PM final concentration) where 15N and 15No are the concentrations of 15NH4+ was added into the container with inflow water and (i.e. total ammonium concentration X atom % 15NH4+) mixed thoroughly. The additions of 10 PM 15NH4+were in the inflow and outflow waters, respectively. Note, selected to provide concentrations of ammonium com- this calculation is based on the assumption that the parable to those produced in the sediments. Some am- NH4+ originating from the sediments or was monium (approldmately 40 % for the example in Table 1) removed at the same relative rate as the added 15NH4+. was removed by organisms in the water before the Finally, the amount of the removed ammonium (Nremoved) water was passed over the cores. Of the water coming on an areal basis (pm01 N m-2 h-') was calculated as: out of the cores, the labeled ammonium constituted 20 to 54 % of the total ammonium. Note, the actual con- centration of dlssolved 15NH4+in the water varied from The concentrations of bacteria, phytoplankton, and day to day because of biological uptake after spiking. protozoa in the inflow and outflow water were mea- After allowing the system to equilibrate over night sured at the beginning and the end of 15NH4+incuba- with the added I5NH4+, we performed 3 sequential tion~.Bacteria preserved with 1 % (final concentration) 15N-incubation intervals and measured ammonium con- formaldehyde were counted on acriflavine-stained centration and atom % "N within each interval to in- filters (Bergstrom et al. 1986).Nanoplankton were pre- crease statistical robustness of the experiment. served with ice-cold glutaraldehyde (1 % final concen- Concentrations of 15NH4+(i.e. NH4+ concentration X tration) and counted using DAPI/FITC double staining atom % "N) were measured in the inflow waters and (Sherr et al. 1993). Microplankton preserved with outflow waters, which were sampled at about the same freshly prepared acid Lugol's iodine (1 % final concen- time. We believe that phytoplankton uptake during tration) were counted in settling chambers. Microbial 190 Aquat Microb Ecol21: 187-194,2000

cell Linear dimensions were measured using an eyepiece micrometer (magnifica- tion up to 1250~)and their biovolumes were calculated by approximating geome- tric solids. Allometric relationships (Nor- land 1993) were used to estimate bacterial from their cell volume and then an average of these estimates was calcu- lated. Protozoan biomass was estimated assuming carbon contents of 0.15 and 0.19 pg C pm-3 for living and preserved cells, respectively (Putt & Stoecker 1989). Algal biovolumes were converted to carbon using the allometric equation for phyto- plankton (Montagnes et al. 1994). o! Under steady state, successive determi- l 2 3 4 5 nations of I5N dark losses and microbial Days parameters were averaged to obtain a sin- Fig. 2. Dynamics of ammonium (-1 and nitrates (- - -) during the initial gle value for each flow-through chamber, phase of the experiment (before the addition of 15NH4+).Mean and standard and then a mean and standard error were error values are calculated for 3 replications at each time point calculated for triplicate experimental treat- ments and control. At the end of the incubation, triplicate samples of fluxes of total ammonium produced by natural rniner- sediments were collected from each core to estimate alization processes than for the 15N-labeled ammonium abundance of benthic protozoa. Benthic protozoa were that had been added on both an absolute and relative counted in several 100 yl subsamples of non-preserved basis (Table 1). sedirnents, which were diluted 1:10 with 0.2 pm The mean percentage of I5NH4+ removed as the filtered lake water (Finlay et al. 1979). Heterotrophic spiked water moved over the sediment cores was nanoflagellates, preserved with 1 % (final concentra- 10 * 2.0 (SE) % for the control treatments as compared tion) glutaraldehyde, were extracted from the sedi- to 26 * 9 % for the zebra mussel treatments. The mean ments using isopycnic centnfugation in shca gel gra- total ammonium flux and removal rate, calculated dients (Epstein 1995) and counted via epifluorescence by Eqs. (2) & (3), respectively, were 68 * 18 and microscopy as described above. 7.3 * 2.5 pm01 N m-' h-' for the control treatment sedirnents as compared to 169 * 30 and 46 * 12 pm01 N m-' h-' for the zebra mussel treatments. RESULTS The biomass of phytoplankton was largely formed by nanoplankton-sized diatoms, cryptophytes, and Chlamy- After some initial oscillation of nutrient concentra- domonas sp. It slightly increased in the control due to tions in control and experimental cores with added the large CymbeUa sp. and Pediastrum sp, that may zebra mussels over the first 24 h, the flow-through sys- tem stabilized (Fig. 2). The concentration of NH4+ in both triplicate outflows remained significantly higher Table 1. Total NH4+ and I5NH,+ (= total concentration times atom % I5N) concentrations (*SE) in inflow water and outflow than in the inflows throughout the experiment, indicat- waters for each time point (T6,T7, and T8) in the 15NH4+-addi- ing a positive flux of NH4+from the sedirnents and ben- tion experiments. All concentrations are expressed as pM N thic organism excretion. Mean net ammonium produc- tion rate (1 17 * 18 [SE], p01N m-' h-') was significantly T6 T7 T8 higher in the zebra mussel treatments than in the Total I5NH4+ Total I5NH4+ Total "NH,+ control treatments (59 * 16 [SE], pm01 N m-' h-'). Conversely, the concentration of nitrates was slightly Inflow 5.9 4.4 6.7 5.7 higher in the inflows than in both outflows, which did Control not differ from each other. Sequential measurements of Mean 10.3 4.2 11.0 5.9 ammonium production rates in individual cores consis- SE 1.2 0.1 1.1 0.2 tently varied less than the mean results obtained from Zebra mussel replicate cores (data not shown). Likewise, variability Mean 15.8 3.7 15.4 4.7 SE 1.6 0.2 1.7 0.4 among cores was greater for the concentrations and Lavrentyev et al.: Cascading effects on N dynamics 191

Phytoplankton zebra mussel cores. However, the proportional con- tribution of the small scuticociliate Uronema sp. was nearly twice as high in the latter treatment, where it comprised ca 60% of the protozoan biomass in the se&ments. In the control, the more diverse protozoan assemblage was reflected in the quantitative presence of large ciliates, such as Spirostomum minus, Loxodes sp., and Oxytricha sp., and a substantial contribution from HNAN. The ciliate Saprodinium sp, occurred in both treatments.

Protozoa DISCUSSION

A steady-state response in our experiments, where ammonium and nitrate fluxes remained stable over & " 8 time, supported the observation of Miller-Way Twil- Inflow Control Zebra Mussel ley (1996) that continuous flow systems can maintain Fig. 3. Mean biomass (*SE)of microorganisms during 15N near-initial experimental conditions throughout incu- incubations bation. Net production and changes in atom % 15N were measurable in all of the cores. Variability of the have been flushed from the sediments. All groups of 15NH4+concentrations and fluxes among cores at a phytoplankton, with the exception of the colonial cy- given time point were consistently lower than those for anobacteriurn Merismopedia sp., drastically decreased the total ammonium concentrations. These results indi- in the zebra mussel treatments (Fig. 3). cate that the ratio of ammonium removal to production In the control, there was no appreciable change in bio- rates in the nepheloid layer was not greatly affected by mass and species composition of the ciliate assemblage, the magnitude of total ammonium concentration or which consisted mainly of oligotrichs and Urotn'cha spp. production rate, at least within the ranges of data that The occurrence of the benthic scuticociliate Uronema sp. were examined here. Net sediment-water ammonium in the outflow water did not significantly affect the total fluxes reported here are similar to those previously ciliate biomass. However, in the treatments with zebra measured in Saginaw Bay (UUmman & Aller 1989). mussels, there was a ca 2-fold decline in protozoan Both ammonium flux and nitrification rate increased biomass. The most dramatic change was observed for significantly in the presence of the zebra mussels. heterotrophic nanoflagellates (HNAN). Their biomass Interestingly, even though total NH4+ fluxes were decreased by 82 % in the zebra mussel treatments. higher, the percentage of the total ammonium flux that Bacterial biomass showed a similar trend to that of was taken up increased by 2.5-fold in the zebra mussel protozoa but in much smaller proportion. The observed treatments as compared to results from the control changes were mainly accounted for by large (>l pm) treatments. Thus, the presence of zebra mussel not and chain-forming rod- and vibrio-shaped cells that only affected absolute ammonium removal rates but represented ca 25% of the total bacterial biomass in also increased the percentage of regenerated ammo- the inflow, >40 % in the control outflow, and <8% in nium that was removed. Increased ammonium removal the zebra mussel outflow. rates partially counterbalanced increased ammonium The ratio of protozoan to bacterial carbon-based bio- production rates in the presence of the zebra mussel. mass decreased from 27.3 * 2.45 (SE) in the control to Positive shifts in the NH4:N03 ratio have been reported 8.72 * 1.39 in the zebra mussel treatments. The ratio of to trigger blooms of non-nitrogen-fixing cyanobacte- HNAN to bacterial biomass declined even stronger, ria, such as Microcystis (Blomqvist et al. 1994). 0.16 * 0.03 in the control versus 0.03 * 0.01 in the zebra Given the drastic decrease in phytoplankton biomass mussel treatments. As mentioned above, the percent- that was observed in the zebra mussel outflow, the age of 15N lost during the passage of water through the increased dark loss of 15NH4+ in this treatment was core treatments was also significantly higher in the likely more the result of enhanced nitrification than of grazed treatments than in the control. algal uptake. The lack of corresponding changes in The total biomass of benthic protozoa (pg C rnl-l), nitrate concentrations can be explained by coupled sampled in the sediments at the end of the experiment, nitrate losses to denitrification that may have occurred did not significantly differ between the treatments, close to the sediment-water interface (Gardner et al. 0.53 k 0.07 (SE) in the control versus 0.70 k 0.08 in the 1987, Henriksen & Kemp 1988). The zebra mussel can 192 Aquat Microb Ecol21: 187-194,2000

stimulate the development of benthic microalgae in organisms in Saginaw Bay (Lavrentyev et al. 1997), fall Saginaw Bay (Lowe & Pilsbury 1995), but we expect within this size category. that N uptake by these organisms may have been Bacterial biomass did not differ significantly be- small during the prolonged dark incubation. The oc- tween the control and the treatments with zebra mus- currence of the benthic ciliate Saprodinium sp. indi- sels, despite a strong decline in protozoan abundance rectly suggests that anoxic conditions conductive to in the latter treatments. Combined with the observed denitrification may have existed in certain regions of changes in bacterial size structure, thls fact implies sediments in this experiment. that mussels may have removed a certain proportion The mechanisms underlying the zebra mussel of the bacterial population from the overlying water. effects on microbial nitrogen transformations in this However, in the latter treatment, a decrease in the pro- study appear to be different from those reported for tozoan to bacterial biomass ratio, possibly reflecting other . In our experiments, the zebra the degree of bacterivory, corresponded to an increase mussels remained either on the top of the sedirnents in the dark 15N loss. This result suggests that a sharp or attached to the side of the core tube above the reduction in HNAN numbers allowed bacterial activity sediment surface. Thus, they were positioned to filter to increase. In Saginaw Bay, the zebra mussel grazed a overlying water rather than pore water. The fact that certain proportion oi the bactenai popuiation, but their the sediments were firm sand rather than clay pre- overall effect on bacterial production in the water col- vented the mussels from penetrating extensively into umn in Saginaw Bay was stimulating (Cotner et al. the sediment matrix. Unlike many other sedirnent- 1995). Protozoa can selectively graze metabolically dwehng macrofauna, which aerate sedirnents via active bacteria (del Georgio et al. 1996, L6pez-Amor6s their bioturbation activities (Pelegri & Blackburn et al. 1998). Although some bivalve mussels are capa- 1995, Rysgaard et al. 1995, Svensson & Leonardson ble of assimilating bacterial food due to their flexible 1996), the zebra mussel does not burrow into the sedi- digestive system (Dech & Luoma 1996), it is unlikely ments, but forms dense beds on solid substrates at the that a similar level of selectivity can occur in them. sediment surface. The potential cascading trophic effect of In Saginaw Bay, the zebra mussel deposits large polymorpha on the microbial community appears to be amounts of and at the sediment similar to those described for planktonic heterotrophic surface (Nalepa et al. 1995) and significantly enhances (Jiirgens & Giide 1994) and nitrifying bacteria (Lavren- N-cycling rates by excreting ammonium (Gardner et tyev et al. 1997). Although we did not measure bac- al. 1995b). However, the concentrations of NH,' were terivory rates in these experiments, our earlier experi- not limiting to microorganisms even in the absence of ments in Saginaw Bay (Lavrentyev et al. 1997)identified zebra mussel, as indicated by the positive net NH,+ HNAN as major consumers of bacteria. Moreover, we flux from the control cores. Under nitrogen-limited found that HNAN were capable of efficiently removing conditions, which are more typical for marine systems, nitrifying bacteria, even when these bacteria were pre- nitrification may be enhanced by NH,+-rich fecal pel- sent at very low concentrations. lets placed on the sediment surface (Henriksen & The effects of bivalve activities appear to be more Kemp 1988). complex in the sediments than in the overlying water. Our results are consistent with the idea that cascad- It is reasonable that the organic-rich feces and pseudo- ing trophic phenomena could directly affect the trans- feces deposited by the zebra mussel at the sediment formation and fate of mineralized nitrogen at the surface could have stimulated the growth of het- sediment-water interface. The zebra mussel filtering erotrophic bacteria in sediments that, in turn, may process appeared to selectively remove and bac- have caused the observed shift toward a predomi- terial grazers from the overlying water. This observa- nance of scuticociliates in microbenthic community. tion corresponds well to the results of earlier studies, These specialized bacterivorous ciliates have much where the mussel drastically reduced the abundance higher specific grazing rates upon bacteria than do of phyto- and protozooplankton (Leach 1993, Lavren- HNAN (e.g. ~imeket al. 1994). However, the lack of tyev et al. 1995, Bastviken et al. 1998, Findlay et al. direct measurements of bacterial biomass in sediments 1998). The size-based prey selectivity of filter-feeding inhibits further speculation at this point. mussels appears to determine their cascading effects Trophic links between microorganisms and bottom- on bacteria. Having a typical filtration rate of ca 50 m1 dwelling invertebrates are not unique to the zebra mussel-' h-' (Reeders & de Vaate 1990), the zebra mus- mussel and the freshwater environment. The estuarine sel is capable of filtering particles from 0.7 to 450 pm bivalve Mytilus edulis affected microzooplankton- (Jerrgensen et al. 1984), but maximum retention occurs sized ciliates (Riemann et al. 1990) and preyed upon in the size range from 5 to 35 pm (Sprung & Rose 1988). bacterivorous heterotrophic flagellates (Kreeger & Heterotrophic nanoflagellates, the major bacterivorous Newell 1996). The oyster Crassostrea gigas has been Lavrentyev et al.- Cascading effects on N dynamlcs 193

recently found to assimilate ciliates and flagellates The manuscript benefited from criticisms provided by 2 (Dupuy et al. 1999). anonymous reviewers. Thls is GLERL contribution 1150 and UTMSI contribution 1128. Combined with these observations, the results of our study suggest a new interpretation of previously published data on the effects of LITERATURE CITED on the nitrification process. For example, in the Chu- kchi Sea, the large bivalve Macoma calcera stimu- Bak RPM, van Duyl FC, Neiuwland G (1995) Organic sedi- mentation and macrofauna as forcing factors in marine ben- lated nitrification despite the fact that it excreted thlc nanoflagellate communities. Mlcrob Ecol 29:173-182 directly to the water column through its exhalant Bastviken DTE, Caraco NF, Cole JJ (1998) Experimental siphon with little irrigation effect in the sediments measurements of zebra mussel (Dreissena polymorpha) (Henriksen et al. 1993). One explanation that agrees mpacts on phytoplankton commumty composition. Freshw Bio139:375-386 with these results is that the bivalves could have Bergstrom I, Heinanen A, Salonen K (1986) Comparison of stimulated nitrification by filtering out bacterivorous acridine orange, acriflavine, bisbenzinlide stains for enu- protozoa that would otherwise graze nitrifying bac- meration of bacteria in clear and humic waters. Appl Env- teria. Iron Microbiol 51:664-667 Another marine study found that sponges positively Binelh A, Provini A, Galassi S (1997) Modifications in Lake Como (N. ) caused by the zebra mussel (Dreissena affected nitrification rates (Diaz & Ward 1997). The dif- polymorpha). Water hrSoil Pollut 99:633-640 ference in accumulation of nitnte and nitrate in the Blomqvist P, Peterson A, Hyenstrand P (1994) Arnrnonium- presence of different species of sponges in that study nitrogen: a key regulatory factor causing dominance of led to the conclusion that bacteria directly associated non-nitrogen-fixing cyanobacteria in aquatic systems. Arch Hydrobiol 132:141-164 with the sponges caused that nitrification. However, Bock E, Koops HP, Harms H (1989) Nitrifying bacteria. In: these data are also interesting with respect to our Brock TD, Schlegel HG (eds) Autotrophic bacteria. Sci- hypothesis regarding cascading trophic effects on the ence Tech, Madison, p 81-96 nitrification process. Filter-feeding sponges could rea- Cotner JB, Gardner WS, Johnson JR, Sada RH, Cavaletto JF, sonably reduce the numbers of bacterivorous protozoa, Heath RT (1995) Effects of zebra mussels (Dreissena poly- morpha) on bacterioplankton: evldence of both size-selec- which could otherwise graze down both suspended tive consumption and growth st~mulation.J Great Lakes and surface-associated nitrifier cells. Res 21:517-528 Although we did not attempt to represent all actual Dame RF, Zingmark RG, Haskin E (1984) Oyster reefs as complex processes at the sediment-water interface in processors of estuarine materials. J Exp Mar Biol Ecol 83: 239-247 our laboratory study, we believe that these data offer Dame RF, Dankers R, Pnns T, Joncsma H, Smaal A (1991) The some preliminary insight into the effects that bivalves influence of mussel beds on nutrients in the Western Wad- such as the zebra mussel may have on nitrogen den Sea and Eastern Scheldt . Estuanes 14: dynamics in nature. The cascading trophic effects 130-138 could be quantitatively important to the transforma- Dech AW, Luoma N (1996) Flexible dgestion strategies and trace metal assimilation in marine bivalves. Limnol Ocea- tions and fluxes of nitrogen in certain sediment-water nogr 41:568-572 environments (e.9, shallow productive areas of lakes, del Giorgio PA, Gas01 JM, Vaque D, Mura P, Agusti S, Duarte estuaries, and the coastal ocean). Conducive features CM (1996) Bacterioplankton community structure: protists include both an environment favorable for active control net production and the proportion of active bacte- ria in a coastal marine community. Lirnnol Oceanogr 41: microbial processes (e.g. nitnfication) and the pres- 1169-1179 ence of dense populations of bivalve mussels or other Diaz MC, Ward BB (1997) Sponge-mediated nltnfication In suspension-feeding invertebrates that can exert 'top- tropical benthic communities. Mar Ecol Prog Ser 156: down' effects on microbial food web components. The 97-107 above example of zebra mussel affecting microbial Dupuy C, Le Gall S, Hartmann HJ, Breret M (1999) Retention of ciliates and flagellat* by the oyster Crassostrea gigas food web composition, as well as the fate of nitrogen at in French Atlantic coastal : protists as a trophlc link the sediment-water interface, points to the need for between bacterioplankton and benthic suspens~on-feed- more research on the effects of cascading trophic ers. Mar Ecol Progr Ser 177:165-175 effects on biogeochemical cycles at both regional and Epstein SS (1995) Simultaneous enumeration of protozoa and micrometazoa from marine sandy sedunents. Aquat mcrob global scales. Ecol9:219-227 Findlay S, Pace ML, Fisher DT (1998) Response of hetero- trophic bacteria to the zebra mussel invasion of the tidal Acknowledgements. Thls study was supported by the GLERL freshwater . Microb Ecol36:131-140 Non-Indigenous Species Program and in part by the Sea Finlay BJ, Laybourn J, Shashan J (1979) A technique for enu- Grant No. NA16RG0457 to Jim Cotner. We thank Jim Cotner, meration of benthic ciliated protozoa. Oecologia 39:375-377 Dave Fanslow, and Tom Johengen for collecting the mussels Gardner WS, St. John PA (1991) High-performance liquid and sedment cores, Tom Johengen for help with nitrate chromatographic method to determine ammonium ion analyses, and Joann Cavaletto for assistance in the laboratory. and primary amines in seawater. Anal Chem 63:537-540 194 Aquat Microb Ecol21: 187-194, 2000

Gardner WS, Nalepa TF, Malczyk JM (1987) Nitrogen miner- Estimating carbon, protein, and chlorophyll a from volume alization and denitrification in sediments. in marine phytoplankton. Limnol Oceanogr 39:1044-1060 Limnol Oceanogr 32:1226-1238 Nalepa TF, Wojcik JA, Fanslow DL, Lang GA (1995) Initial Gardner WS, Seitzinger SP, Malczyk JM (1991) The effects of colonization of the zebra mussel (Dreissena polymorpha) sea salts on the forms of nitrogen released from estuarine in Saginaw Bay, Lake Huron: population recruitment, and freshwater sediments: does ion pairing affect amrno- density, and size structure. J Great Lakes Res 21:417-434 nium flux? Estuaries 14:157-166 Norland S (1993) The relationship between biomass and vol- Gardner WS, Bootsma HA, Evans C, St. John PA (1995a) ume of bacteria. In: Kemp PF, Sherr BF, Sherr EB. Cole JJ Improved chromatographic analysis of l5N:l4N ratios in (eds) Handbook of methods in aquatic microbial ecology. ammonium or nitrate for isotopic addition experiments. Lewis, Boca Raton, p 303-307 Mar Chem 48:2?1-282 Pelegri SP, Blackburn TH (1995) Effect of bioturbation by Gardner WS, Cavaletto JF, Johengen TH, Johnson JR, Heath Nereis sp., Mya arenaria and Cerastoderma sp, on nitrifi- RT, Cotner JB (1995b) Effects of the zebra mussels, Dreis- cation and denitrification in estuarine sediments. Ophelia sena polymorpha, on community nitrogen dynamics in 24:289-299 Saginaw Bay, Lake Huron. J Great Lakes Res 21:529-544 Putt M, Stoecker DK (1989) An experimentally determined Henriksen K. Kemp WM (1988) Nitrification in estuarine and carbon:volume ratio for marine oligotrichous clliates coastal marine sediments. In: Blackburn TH, Ssrensen J from estuarine and coastal waters. Limnol Oceanogr 34: (eds) Nitrogen cycling in coastal marine environments. 1177-1183 Wdey & Sons, C'nicnesier, p 207-240 Reeders HE, dc Vas!r AB (!000) ZeSrs mussels [Drdssena Henriksen K, Blackburn TH, Lomstein BAa, McRoy CP (1993) polymorpha): a new perspective for water quality man- Rates of nitrification, distribution of nitrifying bacteria, agement. Hydrobiologia 200/201:437-450 and inorganic N fluxes in northern Bering-Chukchi shelf Riemann B, Serensen HM, Bjernsen PK, Horsted SJ, Jensen sediments. Cont Shelf Res 13:629-651 LM, Nielsen TG, Sandergaard M (1990) Carbon budgets Jsrgensen CB, Kisrboe T, Riisgard HU (1984) Ciliary and of the microbial food web in estuarine enclosures. Mar mucus-net filter feeding, with special reference to fluid Ecol Prog Ser 65:159-170 mechanical characteristics. Mar Ecol Prog Ser 15:283-292 Rysgaard S, Christensen PB, Nielsen LP (1995) Seasonal vari- Jiirgens K, Giide H (1994) The potential importance of graz- ation in nitrification and denitrification in estuarine sedi- ing-resistant bacteria in planktonic systems: a review. Mar ment colonized by benthic microalgae and bioturbating Ecol Prog Ser 112:169-188 infauna. Mar Ecol Prog Ser 126:lll-121 Kreeger DA, Newell RIE (1996) Ingestion and assimilation of Seitzinger SP (1988) Denitrification in freshwater and coastal carbon from celluolytic bacteria and heterotrophic fla- marine ecosystems: ecological and geochemical signifi- gellates by the mussels Geukensia demissa and Mytdus cance. Limnol Oceanogr 33:702-724 edulis (Bivalva, ). Aquat Microb Ecol 11:205-214 Sherr EB, Caron DA, Sherr BF (1993) Staining of hetero- Lavrentyev PJ, Gardner WS, Cavaletto JF, Beaver JR (1995) trophlc protists for visualization via epifluorescence micro- Effects of the zebra mussel (Dreissena polymorpha Pallas) scopy. In: Kemp PF, Sherr BF, Sherr EB, Cole JJ (eds) on protozoa and phytoplankton in Saqnaw Bay, Lake Handbook of methods in aquatic microbial ecology. Lewis, Huron. J Great Lakes Res 21:545-557 Boca Raton, p 213-228 Lavrentyev PJ, Gardner WS, Johnson JR (1997) Cascading ~imekK, Vrba J, Lavrentyev PJ (1994) Estimates of protozoan trophic effects on aquatic nitrification: experimental evi- bacterivory: from microscopy to ectoenzyme assay? Mar dence and potential implications. Aquat Microb Ecol 13: Microb Food Webs 8:71-85 161-175 Sprung M, Rose U (1988) Influence of food size and food qual- Leach JH (1993) Impacts of the zebra mussel (Dreissena poly- ity on the feedmg of the mussel Dreissena polymorpha. morpha) on water quality and spawning reefs in west- Oecologia 77:526-532 ern . In: Nalepa TF, Schloesser DW (eds) Zebra Strickland JDH, Parsons TR (1972) A practical handbook for mussels biology, impacts, and control. Lewis, Ann Arbor, seawater analysis, 2nd edn. Bull Fish Res Board Can 167 p 381-388 Suzuki M, Sherr EB, Sherr BF (1996) Estimation of ammonium Lopez-Amorbs R, Comas J, Garcia MT, Vives-Rego J (1998) regeneration efficiencies associated with bacterivory in Use of the 5-cyano-2,3-ditolyl tetrazolium chloride reduc- pelagic food webs via a 15N tracer method. J Plankton Res tion test to assess respiring marine bacteria and grazing 18:411-428 effects by flow cytometry during linear alkylbenzene sul- Svensson JM, Leonardson L (1996) Effects of bioturbation by fonate degradation. FEMS Microbial Ecol27:33-42 tube-dwelling chironomid larvae on oxygen uptake and Lowe RL, Pillsbury RW (1995) Shifts in benthic algal commu- denitrification in eutrophic lake sediments. Freshw Biol nity structure and function following the appearance of 35:289-300 zebra mussels (Dreissena polymorpha) in Sagnaw Bay, Ullman WJ, Aller RC (1989) Nutrient release rates from the Lake Huron. J Great Lakes Res 21:558-566 sediments of Saginaw Bay, Lake Huron. Hydrobiologia Miller-Way T, Twilley RR (1996) Theory and operation of con- 171:12?-140 tinuous flow systems for the study of benthic-pelagic cou- Ward BB (1986) Nitrification in marine environment. In: pling. Mar Ecol Prog Ser 140:257-269 Prosser JI (ed) Nitrification. Oxford Univ. Press, Oxford, Montagnes DJ, Bergers JA, Harrison PJ, Taylor FJR (1994) p 157-184

Editorial responsibility: Patricia Glibert, Submitted: May 11, 1999; Accepted: October 25, 1999 Cambridge, Maryland, USA Proofs received from author(s): February 29, 2000