Bioturbation Effects of Chironomus Riparius on the Benthic N-Cycle As Measured Using Microsensors and Microbiological Assays

AQUATIC MICROBIAL ECOLOGY Vol. 27: 175–185, 2002 Published March 15 Aquat Microb Ecol

Bioturbation effects of Chironomus riparius on the benthic N-cycle as measured using microsensors and microbiological assays

Peter Stief 1, 2,*, Dirk de Beer 2

1Department of General Ecology and Limnology, University of Cologne, Weyertal 119, 50923 Köln, Germany 2Max Planck Institute for Marine Microbiology, Celsiusstr. 1, 28359 Bremen, Germany

ABSTRACT: Chironomus riparius (Diptera) larvae were added to laboratory microcosms containing defaunated sediments sampled at 2 NO3-polluted field sites. Following a 3 wk incubation, the larval – + influence on the sedimentary nitrogen conversions was studied using microsensors (O2, NO3 , NH4 ) and microbial bulk parameters (microbial biomass, community respiration). At the sediment surface the chironomid larvae fed on particles (deposit-feeding layer), while in the subsurface zone the larvae moved through the sediment and ventilated transient or permanent burrows (ventilation layer). – + In the deposit-feeding layer of the chironomids, NO3 production and NH4 consumption were lower – + and microbial biomass decreased. In the ventilation layer of the chironomids, NO3 and NH4 conversion maxima were shifted downwards, and both microbial biomass and community respiration were increased. The observed changes in the vertical stratigraphy of the benthic microbial community were ascribed to the depth-specific larval behaviour as: (1) particle ingestion and removal of adhering microorganisms in the deposit-feeding layer; and (2) stimulation of subsurface micro- – organisms due to an increased supply of O2 and NO3 along with larval ventilation.

KEY WORDS: Freshwater sediment · Nitrogen cycle · Chironomus riparius · Bioturbation · Microsensor · Microbial biomass · Community respiration

Resale or republication not permitted without written consent of the publisher

INTRODUCTION supply of O2 from the water column to the sediment + and to the excretion of NH4 by animals (Mayer et al. – Against the background of the continual NO3 pollu- 1995). The stimulatory effect on denitrification has tion of surface waters in Central Europe (Hellmann been ascribed to the increased sediment/water inter- 1992), sediments of aquatic ecosystems have received face area (Pelegri & Blackburn 1995), to the additional – – considerable attention as possible NO3 sources and NO3 supply from the water column (Svensson 1997, sinks (Bowden 1986, Howard-Williams & Downes Gilbert et al. 1998) and to a tighter coupling of nitrifi- 1993). The benthic macrofauna stimulates both nitrification and denitrification (Pelegri & Blackburn 1994, – cation, i.e. the microbial production of NO3 , (Mayer et Svensson 1998, Tuominen et al. 1999). al. 1995) and denitrification, i.e. the microbial con- In a previous study, the enhancement of nitrification – sumption of NO3 (Kristensen et al. 1985, Pelegri & in the presence of bioturbating animals was assessed – Blackburn 1994, 1995, Svensson 1997, 1998, Gilbert et from slurry incubations and specific inhibition of NO2 al. 1998, Tuominen et al. 1999). The stimulatory effect oxidation (Mayer et al. 1995). The animal-induced on nitrification has been ascribed to the increased stimulation of denitrification has been quantified by 15 – determining the fate of NO3 enrichments in the over- *Address for correspondence: Max Planck Institute for lying water of bioturbated sediments (Pelegri et al. *Marine Microbiology. *E-mail: [email protected] 1994, Pelegri & Blackburn 1995, Svensson 1997, 1998).

The vertical distribution of nitrogenous solutes has sampled from the Liblarer Mühlengraben (LMG), a been recorded in the pore-water on the scale of a few man-made brook near Cologne, Germany. Both field millimetres or centimetres (Pelegri & Blackburn 1994, sites are NO3-polluted, either due to groundwater Tuominen et al. 1999). The low spatial resolution of seepage through the sediments (GAR) or drainage these concentration profiles prevents calculation of water from farmland and effluents from a municipal – local NO3 consumption rates in the microenvironment wastewater treatment plant (LMG). of burrowing animals. The destructive nature of pore- The Chironomus riparius (Meigen) larvae were water extraction did not allow a single sample to be taken from a laboratory bred population originating analysed with respect to both its content of nitroge- from the silty deposits of LMG, where they occurred in nous solutes and to its microbial characteristics. These densities of up to several 10s of 1000s of individuals per limitations can be overcome by using ion selective m2 during the summer of 1995. The larvae were reared microsensors in combination with small scale applica- in aquaria filled with a thin layer of ground dolomite tions of microbial bulk parameters. and an aerated synthetic freshwater medium (Stief & In this study of bioturbation, we hypothesised that Neumann 1998). Larvae were regularly fed with a sus- the presence of Chironomus riparius larvae can either pension of shredded leaves of Urtica sp. and exposed suppress or stimulate the benthic microbial community to 15°C and a light:dark cycle of 12:12 h. depending on the sediment depth of consideration: At Microcosm preparation and incubation. The sedi- the sediment surface, the larvae deposit feed and ments were passed through a 1 mm mesh to remove thereby reduce the microbial biomass associated with pebbles, large detritus particles and indigenous the sediment and detritus particles (Johnson et al. macrofauna. After thorough mixing they were appor- 1989, van de Bund et al. 1994). Larval grazing at the tioned into cylindrical Perspex sediment containers of sediment surface might thus lower the abundance and 2 sizes (9.5 and 19 cm in diameter, both 13 cm high) to activity of microorganisms involved in the nitrogen give a final sediment height of 10 cm. The compacting cycle. In subsurface sediment layers, however, the lar- sediments were covered by aerated synthetic freshwa- vae construct and ventilate transient or permanent ter (Stief & Neumann 1998). Sediment microcosms burrows creating additional surfaces for microbial were allowed to stabilise for 2 wk at 15°C and a – colonisation which are supplied with O2 and NO3 from light:dark cycle of 12:12 h without further manipula- the overlying water (Kristensen et al. 1985, Mayer et al. tions. 1995, Pelegri & Blackburn 1995, Svensson & Leonard- For experiments, a known amount of 4th instar son 1996, Svensson 1997, 1998). Larval ventilation of larvae was introduced into the sediment microcosms subsurface burrow structures might thus increase the (0 or 1 ind. cm–2). During the experimental incubation abundance and activity of microorganisms involved in no Urtica sp. suspension was added, meaning that the the nitrogen cycle. sediment and its microbial community served as the The stimulatory or suppressive effects of chironomid only food source for the larvae. The water column bioturbation on the distribution and activity of the ben- above the sediments was supplied with 15 ml h–1 of –1 – thic microbial community were studied in replicate synthetic freshwater containing 475 µmol l NO3 . This sediment microcosms, which were either devoid of or corresponded to dilution rates of 0.08 and 0.02 h–1 in inhabited by Chironomus riparius larvae. Microbial the small and large sediment containers, respectively. biomass and community respiration were quantified in The microcosms were inspected daily for dead larvae sediment slices of 2 mm thickness. Pore-water concen- or emerged midges that were replaced by new larvae. – + trations and conversion rates of O2, NO3 and NH4 After a 3 wk incubation, all analytical procedures were were measured using microsensors. Our hypothesis completed within 6 d. Subsequently, the sediment was tested for organic-poor sand versus organic-rich cores were sieved in order to retrieve and quantify the silt, taken from 2 NO3-polluted habitats of chirono- larvae still alive. mids. Experimental design. Six sediment microcosms were run each with organic-poor sediment (site GAR, large sediment containers) and organic-rich sediment MATERIALS AND METHODS (site LMG, small sediment containers). Three microcosms of each sediment type served as control treat- Origin of sediments and chironomids. Surface sediments to which no chironomid larvae were added, ment (0 to 15 cm) was collected from 2 contrasting field while the other 3 microcosms served as experimental sites that contained chironomids. Sandy deposits were treatments with 1 Chironomus riparius larva cm–2 each. sampled in the littoral zone of the Grietherorter Microsensor measurements (microbial nitrogen con- Altrhein (GAR), a disconnected oxbow of the Lower versions) were made between Days 18 and 24 of the River Rhine, Germany, and silty sediment material was incubation at randomly chosen spots of the micro- Stief & de Beer: Chironomus riparius and the benthic N-cycle 177

cosms. Extractable ATP content (microbial biomass important features of the profiles, but led to a slight estimate) and INT reduction capacity (respiratory smearing of production-consumption zones. activity estimate) were analysed in sliced sediment The depth-integrated flux Jdi of a solute was calcu- subcores between Days 21 and 25 of incubation after lated as the sum of all R values multiplied by the thick- the microsensor measurements had been completed in ness of a single conversion zone: a microcosm. J = (ΣR )(x – x ) (5) Physico-chemical sediment characteristics were di i i+1 i either determined using subsamples of the homo- (from i = 1 to i = the maximum number of conversion + genised source material (for organic content, protein zones). In the case of NH4 , the depth-integrated flux content, and grain-size distribution) or observed was determined for the anoxic zone of the sediment through the container wall at the end of the sediment column rather than across the sediment/water inter- incubation (thickness of oxidised layer). face. For this purpose the linear part of the concentra- Microsensor measurements and data interpretation profile below the oxic/anoxic interface was used – + tion. LIX-type microsensors, selective for NO3 and together with Eq. (1) to calculate the NH4 release into + NH4 , and Clark-type O2 microsensors were con- the oxic zone (de Beer et al. 1991). structed as have been described (de Beer & van den The microsensor-based fluxes (henceforth referred

Heuvel 1988, Revsbech 1989, Sweerts & de Beer 1989). to as depth-integrated fluxes, Jdi) were cross-checked Individual microsensors were calibrated in synthetic using the solute concentration in the continuously freshwater (Stief & Neumann 1998) with known diluted overlying water. These overlying water-based amounts of the respective ions. Diffusion coefficients of fluxes (henceforth referred to as overlying water – + –5 NO3 , NH4 and O2 at 15°C were taken as 1.44 × 10 , fluxes, Jow) can be calculated as: 1.50 × 10–5 and 1.83 × 10–5 cm2 s–1, respectively. The J = (C – C )V/A (6) microsensors were driven into the sediments by a com- ow ow dil puter controlled micromanipulator at step sizes of where Cow is the solute concentration in the overlying between 100 and 1000 µm. Vertical pore-water micro- water, Cdil is the solute concentration in the supplied profiling was repeated 1 to 3× in each sediment micro- synthetic freshwater, V is the flow rate, and A is the cosm. During these measurements the water column cross-sectional area of the sediment. was supplied with new synthetic freshwater at a Microbial biomass and community respiration. known dilution rate. The mixing of the water column Microbial biomass: The extractable amount of ATP and the maintainance of a diffusive boundary layer was determined for the layers 0 to 2, 2 to 6 and 6 to above the sediment surface was accomplished by 10 mm using a modified version of the protocols of Karl sparging the overlying water with air. & LaRock (1975) and Karl & Craven (1980). A sediment The concentration profiles were converted into a volume of 1 cm3 (devoid of larvae) was suspended in –1 vertical sequence of local conversion rates applying an 5 ml of the ice cold extractant (48 mmol l EDTA-Na2 extension of Fick’s law of diffusion: in 1 mol l–1 phosphoric acid) and stored on ice for 30 min. After shaking once again, the suspension was J = ϕD(δC/δx) (1) centrifuged and the supernatant diluted 1:50 with where J is the flux, ϕ is the sediment porosity, D is the sterile deionised water and its pH adjusted to 7.8 with diffusion coefficient, and δC is the concentration gradi- NaOH. A luciferase assay kit (Sigma Chemical) and ent along the distance δx. Local fluxes at the exem- the luminometer TD 20 (Turner Designs) were used to plary chosen depths 1 and 2 were approximated as: perform the firefly bioluminescence reaction. After addition of known amounts of cellular-free ATP to J = ϕ D(C – C )/(x – x ) (2) 1 1 0 2 0 2 sediment samples, an ATP loss of 15% was noted. All measured results were thus corrected for this apparent J = ϕ D(C – C )/(x – x ) (3) 2 2 1 3 1 3 loss. The volumetric conversion rate within the layer Community respiration: The capacity of the electron defined by the depths 1 and 2 was then calculated as: transport system (ETS) of the benthic microbial community was determined through the reduction of INT R = (J – J )/(x – x ) (4) 1, 2 1 2 1 2 to INT-formazan (Blenkinsopp & Lock 1990). A sedi- Two consecutive layers were lumped together ment volume of 1 cm3 from the layers 0 to 2, 2 to 6 and according to (R1 + R2)/2 = R1', (R3 + R4)/2 = R2', etc. and 6 to 10 mm (devoid of larvae) was incubated with 5 ml afterwards the weighed running average of 3 consecu- of 0.02% INT for 2 h at 15°C in centrifuge tubes. Non- tive layers was calculated according to (0.5 × R0' + R1' + converted INT was removed by centrifuging, dis- 0.5 × R2')/2 = R1'', (0.5 × R1' + R2' + 0.5 × R3')/2 = R2'', etc. carding the supernatant and washing the pellet with These 2 smoothing procedures helped highlight the filter-sterilised synthetic freshwater (repeated 3×). 178 Aquat Microb Ecol 27: 175–185, 2002

INT-formazan was extracted from the remaining pellet Table 2. Depth of redox discontinuity. Depth range of the with 98% methanol for 1 h at 4°C and sonication for colour change of the sediment from light-brown to dark- 5 min. The extract was centrifuged again and the ex- brown (organic-poor) or black (organic-rich) is given tinction of the supernatant was determined at 480 nm. The extinction values of formaline-killed replicate Sediment type Depth of redox discontinuity (mm) samples were subtracted from those of the living sam- Without Chironomus With Chironomus ples. INT-formazan formation rates were converted to Organic-poor 9–10 10–13 volumetric O2 consumption rates (Relexans 1996). Organic-rich 50 5–7 Physico-chemical sediment characteristics. Organic and protein content: Approximately 20 ml of each of the homogenised sediments were poured into precom- change of sediment colour moved deeper down into busted glass beakers and dried to constant dry weight. the sediment in the presence of chironomid larvae The dried samples were combusted at 550°C for 3 h. (Table 2). In the non-bioturbated sediments the light- Organic content was estimated as weight loss during this brown oxidised layer had an invariable thickness, process. The extractable protein content was determined while in the bioturbated sediments a few oxidised according to Rausch (1981) using 150 to 1000 mg of dried halos were observed around chironomid burrows. sediment of the homogenised source material. Both analytical procedures were replicated 4 to 10 times. Thickness of oxidised layer: Redox-dependent Larvae colour change in the sediment column was observed through the transparent container wall at the end of Larval burrowing activity in the 2 sediment types the incubations. The light-brown top layer of the sedi- was identical in that fewer than 10% of the added ments was defined as the oxidised layer. chironomids constructed a permanent, U-shaped bur- Mean grain diameter: Approximately 250 ml of wet row reaching down to 10 mm. Other individuals moved sediment was sieved through a set of analytical sieves within the subsurface zone of the sediments at depths with the mesh sizes 63, 160, 200, 250, 400 and 630 µm. of between 2 and 6 mm. Ventilation activity of the

The mean grain diameter d50 was determined from the larvae in the subsurface zone was sometimes detected grain-size distribution curve. as rhythmic up and down movements of the sediment surface. The larvae collected and ingested sediment or detritus particles at the sediment surface. A small RESULTS amount of shredded Urtica sp. leaves placed around a burrow outlet was picked up by the larvae within a few Sediments and larvae minutes. Chironomid locomotive and feeding activity made the sediment surfaces look fluffier than occurred Sediments in the non-bioturbated control microcosms. More adult midges emerged during the experimental incubation Details on organic content, protein content and mean in the organic-rich than in the organic-poor sediments grain size are given in Table 1. The sediment collected (40 and 25%, respectively). Up to 35 and 50% of the from the GAR was a light-brown, organic-poor sand added larvae were found alive at the end of the incu- with a low content of visible detritus. Protein could be bation in the organic-rich and the organic-poor sedi- extracted only in trace amounts from this sediment ments, respectively. The larvae that neither emerged relative to the dark-brown, organic-rich silt taken from nor were found alive were assumed to have died. Lar- the LMG. In both sediment types the redox-dependent val mortality accounted for up to 25% of the total addi- tions in both sediment types. Only a few dead individuals were retrieved; the remaining larvae had Table 1. Initial physico-chemical sediment characteristics. decomposed. Means (±SD, number of repeated analyses) are given

Homogenised source material Microbial oxygen and nitrogen conversions Organic-poor Organic-rich sediment sediment General observations

Organic content (mg g–1) 12.9 (±1.6, 8) 43.8 (±1.9, 10) The concentration profiles in Fig. 1A–F represent Protein content (mg g–1) 1.1 (±0.1, 4) 18.2 (±0.7, 4) Mean grain diameter (µm) 306 (±12, 4) 108 (±5, 4) the steady state distribution of solutes after 3 wk incubations. Their smooth curvature indicates continuous Stief & de Beer: Chironomus riparius and the benthic N-cycle 179

producing and the NO3-consuming zones (Fig. 3B,E). In the organic-poor sediment, the maximum local conversion rates remained unchanged after the 3 wk colonisation with chironomids. In the organic-rich sediment, however, – the local NO3 production and consumption rates were lower when animals were present. Integrated over depth, the presence of chironomid larvae decreased (in organic-poor sediment) or increased (in organic-rich – sediment) the sedimentary NO3 consumption (Fig. 4B,E), but these differences were not statistically significant when based on microsensor measure-

ments (i.e. depth-integrated fluxes, Jdi). In contrast to this result, the overlying

water fluxes, Jow, indicated significant – differences in the sedimentary NO3 Fig. 1. Vertical concentration profiles in sediment microcosms without between the bioturbated and the non- –2 e –2 f Chironomus riparius (0 larvae cm , ) or with C. riparius (1 larva cm , ). bioturbated sediment (Fig. 4C,F). Means (±SD) of 3 microcosm replicates with 1 to 3 repeated profiles each are shown. ‘poor’ = organic-poor sediment, ‘rich’ = organic-rich sediment

+ NH4 conversions profiles, uninterrupted by the unique microenvironment of animal burrows. In Fig. 2A,B,C it can be seen, The 3 wk bioturbation activity by Chironomus – + + however, that pore-water NO3 and NH4 concentra- riparius erased the NH4 concentration gradients + tions approximated bulk-water values at a relatively (Fig. 1C,F), resulting in lower local NH4 consumption + great sediment depth. Irregular curvatures of this type rates near the sediment surface and lower local NH4 occurred in 3 out of 81 profile recordings and will be production rates deep in the sediment (Fig. 3C,F). + referred to as burrow crossings. Based on the microprofiles presented here, an NH4 flux across the sediment/water interface was not + detected in either direction. Conversely, an NH4 flux O2 conversions from the anoxic zone up into the oxic zone of the sediment could be calculated from the steepness of the In the organic-poor sediment, the presence of larvae concentration gradients in the anoxic zone (Fig. 4C,F). increased local O2 consumption rates near the sedi- The 3 wk presence of C. riparius larvae led to lower + ment surface (Fig. 3A), while in the organic-rich sedi- NH4 fluxes up into the oxic layer, but this trend was ment these rates were lower near the sediment surface only significant in the organic-rich sediment. and higher in the subsurface layer (Fig. 3D). In a depth-integrated budget of the organic-poor sediment, an animal- related increase of sedimentary O2 consumption was seen (Fig. 4A). The organic- rich sediment, in contrast, consumed less

O2 in the presence of chironomids (Fig. 4D).

– NO3 conversions

Fig. 2. Vertical concentration profiles in bioturbated microcosms (1 larva Bioturbation generally caused a greater cm–2), as measured with microsensors at randomly chosen spots of the sedi- – NO3 penetration depth (Fig. 1B,E) and a ment cores. Profiles exhibiting conspicuous irregularities of their curvature downward relocation of both the NO3- are shown 180 Aquat Microb Ecol 27: 175–185, 2002

Microbial bulk parameters

Microbial biomass

In all incubated sediments, microbial biomass was greatest at the sediment surface and then decreased with depth (Fig. 5A,B). The 3 wk presence of chironomid larvae levelled off these vertical microbial biomass gradients. At the end of the 3 wk incubation with Chironomus riparius larvae, the microbial biomass of the 0 to 2 mm layer was reduced in both sediment types tested. Within the 2 to 6 and 6 to 10 mm layers, however, the presence of chironomid larvae caused an amplified microbial biomass.

– + Community respiration Fig. 3. Local production and consumption rates of O2, NO3 and NH4 as derived from concentration profiles. Non-bioturbated cores (open bars), bioturbated cores (closed bars). Positive values correspond to production and negative In the absence of chironomid larvae values to consumption. Means (±SD) of 3 microcosm replicates with 1 to 3 the respiratory activity of the microbial repeated profiles each are shown community decreased at greater sediment depths (Fig. 5C). However, after colonisation of the sediments with

– Fig. 4. Fluxes of O2 (oxic zone) NO3 (oxic and anoxic zone), + and NH4 ( anoxic zone), as derived from punctiform concentration profiles (= depth-integrated fluxes, Jdi: A,B,D,E) or from overlying water concentrations (= overlying water fluxes, Jow: C,F). Non-bioturbated cores (open columns), bio- Fig. 5. Vertical distribution of microbial biomass (A,B) and turbated cores (closed columns). Positive values correspond to microbial community respiration (C), as analysed in sliced an efflux and negative values to an influx. Means (±SD) of 3 subcores taken from sediment microcosms. Non-bioturbated microcosm replicates with 1 to 3 repeated profiles each are cores (open columns), bioturbated cores (closed columns). shown. Student’s t-tests between experimental and control Means (±SD) of 3 microcosm replicates are shown. Student’s treatments revealed significant differences at p < 0.01 (##) t-tests between experimental and control treatments revealed and p < 0.001 (###) significant differences at p < 0.01 (##) and p < 0.001 (###) Stief & de Beer: Chironomus riparius and the benthic N-cycle 181

Chironomus riparius larvae, conspicuous activity of attached microorganisms could have reduced the maxima appeared within the 2 to 6 and 6 to 10 mm microbial biomass (Johnson et al. 1989), but pore- layers. These elevations over background activities water ventilation and larval secretion of organic or were statistically significant, but no change was nitrogenous compounds might have stimulated the observed within the 0 to 2 mm layer. microbial respiratory activity (Svensson & Leonardson 1996).

DISCUSSION – NO3 conversions in the organic-poor sediment Microbial oxygen and nitrogen conversions – NO3 production and consumption zones moved O2 conversions deeper down into organic-poor sediments when chironomids were present. This observation was true even

In the organic-poor sediment, the increase of the though the diminished O2 penetration depth may have –2 –1 – sedimentary O2 consumption (0.23 mmol m h ) was caused the opposite result: the restriction of the NO3 in the range of the presumable larval respiration rate production zone (the presumable nitrification zone) to as recalculated per unit area of sediment (0.19 mmol the thinner layer of intense O2 consumption and the re- –2 –1 – m h , Bairlein 1989). Since some of the larval O2 con- location of the NO3 consumption zone (the presum- sumption must have occurred during the time the lar- able denitrification zone) in the direction of the vae had spent on top of the sediments, the actual larval oxic/anoxic interface. In spite of this prediction, the ob- respiration rate per unit area of sediment would be served downward shift of the nitrification zone can be –2 –1 + lower than the expected 0.19 mmol m h . Thus, an explained by the depletion of NH4 following the long- unknown proportion of the increased sedimentary O2 term presence of Chironomus riparius. This depletion consumption could be attributed to an increased must have restricted the nitrification activity to a + microbial community respiration. For high densities of deeper layer in the sediment where NH4 became Tubifex tubifex (Oligochaeta) a similar result was available. Trace amounts of O2 could be measured explained by the superficial accumulation of organic down to the depth of the observed maximum nitrifica- particles including the associated microorganisms tion rate. The portion of nitrification activity below the

(Pelegri & Blackburn 1995). Indeed, in the top layer of O2 penetration depth may be due to smearing from our sediments a higher content of combustible organic smoothing the local conversion rates. Denitrification matter was found in the bioturbated than in the non- might have become substrate limited from the continu- bioturbated treatments (7.7 ± 0.8% instead of 5.7 ± ous chironomid bioturbation as well: The animal-in- 0.6%, means ± SE, from a total of 8 samples from 3 duced accumulation of organic matter in the top layer replicate treatments each). The higher organic content of the sediments occurred at the expense of deeper could have supported the growth of heterotrophic sediment layers (1.7 ± 0.2% instead of 2.4 ± 0.4%, microorganisms resulting in an increased sedimentary means ± SE, from a total of 8 samples from 3 replicate

O2 consumption. An animal-related increase of super- treatments each) and this transfer might have created ficial nitrifying activity could explain the increased a lack of electron donors needed for denitrification. – sedimentary O2 consumption as well, but the NO3 and Following these ideas, O2 was consumed in a thin top + – NH4 microprofiles demonstrate that this was not the layer of accumulated organic matter, while NO3 was case (Fig. 3B,C). reduced less efficiently due to a lack of suitable or-

In the organic-rich sediment, the sedimentary O2 ganic compounds in deeper sediment layers. The chi- consumption decreased by the end of the 3 wk pres- ronomid bioturbation shifted both the nitrification and ence of Chironomus riparius. Similarly, bioturbation denitrification zone towards sediment layers in which activity of the amphipod Monoporeia affinis caused a the respective substrate demands were still met. – progressive aeration of an organic-rich sediment The sedimentary NO3 consumption in the presence (Tuominen et al. 1999). Both observations could be due or absence of Chironomus riparius showed identical to the loss of microbial biomass close to the sediment trends, no matter if calculated from microprofiles or surface as we observed using ATP extractions. The bulk-water samples. The latter method largely microbial community respiration in the same sediment reduced the scatter of data because the lateral hetero- layer, however, was not different in either the presence geneities, as revealed by the microprofiles taken at or absence of animals. This apparent contradiction can separate spots of the sediment surface, were inte- be resolved when one assumes that bioturbation grated to form a single value. Similar to earlier effects on the distribution and the activity of the micro- attempts to quantify the impact of burrowing animals – bial community cancel each other out: Larval grazing on the benthic NO3 budget, the bulk-water samples 182 Aquat Microb Ecol 27: 175–185, 2002

revealed significant differences between bioturbated tion potentials within the burrow walls of various and non-bioturbated sediment cores (Pelegri et al. marine animal species (Mayer et al. 1995). The inter-

1994, Pelegri & Blackburn 1995, Svensson 1997, 1998). mittent transport of O2-rich water into the burrows + However, microsensors were needed to demonstrate together with the availability of NH4 are believed to the way these differences occurred (i.e. through verti- create favourable growth conditions for nitrifiers. In + cal shifts of production and consumption zones). our case, however, the local NH4 consumption rates close to the sediment surface were lower in the presence of C. riparius. This observation coincides with a – NO3 conversions in the organic-rich sediment lower extractable ATP content in the surface layer of our bioturbated sediments. Larval ingestion and diges- The chironomid larvae helped create a smoother tion of particles and attached microorganisms is one – curvature of the NO3 profiles in the organic-rich sedi- possible way of lowering the abundance of, e.g. nitrify- ments that resulted in lower local conversion rates of ing bacteria. It has been shown for Chironomus plumo- – both NO3 production and consumption. Since vertical sus that larval and bacterial abundances were nega- relocations of the respective conversion zones were of tively correlated to each other after a 10 d incubation in minor significance, the substrate demands of nitrifica- laboratory sediments (Johnson et al. 1989). Similarly, tion and denitrification were met at identical sediment the 22 d bioturbation activity of C. riparius larvae at depths regardless of whether Chironomus riparius abundances of 5000 and 10 000 ind. m–2 led to a larvae were present or not. From the lower local con- decrease of bacterial abundance in the oxidised sur- + version rates, however, a relative depletion of NH4 (as face layer of the sediments (van de Bund et al. 1994). In a substrate of nitrification) and organic compounds (as both cases, larval grazing of attached bacteria was substrates of denitrification) could be inferred. The considered to be responsible for the decrease in bacte- shortage of these compounds at the end of the sedi- rial abundance close to the sediment surface. + ment-animal incubation may be a result of the contin- The NH4 concentration gradient below the oxic/ uous pore-water flushing through larval ventilation anoxic interface within the sediment column was cho- + (i.e. a loss of dissolved compounds to the overlying sen to quantify the flux of NH4 into the presumable water through advective mass transfer). Animals may nitrification zone (de Beer et al. 1991). In the presence + have also enhanced or stimulated the microbial con- of chironomid larvae, less NH4 was supplied from the version of these compounds at the beginning or during anoxic zone to the nitrification zone than in the non- the whole incubation period. Neither one of these pro- bioturbated treatment. Similar observations have been cesses were actually quantified, but they could have made for Chironomus plumosus in that at the end of a + created the substrate-depleted situation at the end of laboratory incubation, interstitial NH4 concentrations the incubation. were lower when animals were present (Fukuhara & – A significant increase of sedimentary NO3 con- Sakamoto 1987). These authors concluded that + sumption was seen in the presence of chironomid mechanical bioturbation led to a gradual loss of NH4 + larvae in depth-integrated budgets. This result was to the overlying water and that NH4 excretion by the obtained by both microprofiling and bulk-water sam- animals could not compensate for this loss. pling. While the latter method delivered statistically significant results (see above), only the microsensors helped elucidate the underlying mechanisms (alter- Microbial bulk parameters – ations of the vertical sequence of local NO3 production and consumption rates). Biomass gradients

Animals with a burrowing behaviour similar to that + NH4 conversions of Chironomus riparius affect the spatial distribution of microorganisms by grazing (Johnson et al. 1989), parti- + Irrespective of the sediment type, the NH4 gradients cle translocation (Francois et al. 1997) or growth stimu- were less steep after the 3 wk incubation with Chi- lation through ventilatory substrate supply (Svensson + ronomus riparius. Both the enhanced transport of NH4 & Leonardson 1996). In our experiments with C. ripar- to the overlying water due to pore-water flushing and ius, the significant decrease of the microbial biomass + the enhanced microbial conversion of NH4 within the within the 0 to 2 mm layer could be ascribed to the sediments due to stimulation of nitrification could deposit-feeding at the sediment surface (Johnson et al. + explain the relative NH4 depletion at the end of the 1989, van de Bund et al. 1994). In contrast, the greater incubation. Experimental support for the latter hypo- microbial biomass in the main residence depth of the thesis has been found in the form of elevated nitrifica- larvae (2 to 6 mm) could have been caused by the Stief & de Beer: Chironomus riparius and the benthic N-cycle 183

stimulation of microbial growth through the supply of rous feeding creates open surfaces that can be – substrates from the overlying water (e.g. O2 or NO3 ). recolonised by microbial populations growing in log Particle translocation could have contributed to the phase. Grazing pressure does not necessarily come flattening of the vertical biomass gradients as well. from macrobenthic organisms; it may also originate The decreased microbial biomass at the sediment from bacterivorous meiofauna (Reichardt 1988) or pro- + – surface coincided with lower NH4 and NO3 conver- tozoa, for which the macrobenthic organisms were cre- sion rates, while in deeper sediment layers the ating favourable growth conditions (Alongi 1985). increased microbial biomass coincided with downward Sieving our sediments through a 1 mm mesh should + – relocated NH4 and NO3 conversion rate maxima. have preserved both groups of animals. Small nema- Even though these observations suggest causal corre- todes were occasionally seen to move within the 0 to lations, one should be careful to bring together a bulk 2 mm layer, but no effort was made to find protozoa. parameter that integrates functional groups of microor- When assuming that apart from the added macro- ganisms with the activities of a rather specific selection invertebrates (Chironomus riparius), bacterivorous of microorganisms. Identification and quantification of meiofaunal species and protozoa were also present in functional groups of microorganisms are needed to our sediments, another trophic cascade seems possi- correlate abundance and activities of microorganisms ble: Chironomids could have grazed upon these small involved in the benthic nitrogen cycle. animals and thereby indirectly decreased the grazing pressure on microorganisms which in return should show higher abundances and activities. A similar Respiration gradients scenario was indeed proposed by Lavrentyev et al. (2000) who showed that the bivalve Dreissena poly- Noticeable elevations of community respiration were morpha removed a large proportion of protozoa (but only seen in the subsurface layers of the bioturbated not bacteria) and thereby indirectly stimulated micro- + sediment and not at the sediment surface. Chironomid bial NH4 conversions at the sediment/water interface. larvae facilitate the transport of solutes from the over- When applied to our experiments, the meiofaunal and lying water down to subsurface microorganisms which protozoan abundances should have been diminished are constrained by a diffusion-limited substrate supply. mainly in the deposit-feeding layer of C. riparius. The – The additional availability of, e.g. O2 and NO3 , proba- expected decrease in bacterivory should then have bly increased the proportion of actively respiring increased the bacterial abundances and activities, but microorganisms, i.e. microorganisms with an operating this could not be confirmed: Microbial biomass was electron transport system (Relexans 1996). Nitrifying smaller, community respiration remained unchanged bacteria, for example, could have profited from O2 in and activities of microbial conversions were mostly the ventilation currents, but because in our sediments lower in the deposit-feeding layer of the chironomids. + NH4 had been depleted at the sediment surface, an activation of nitrifiers was only possible in subsurface layers (e.g. in the 2 to 6 mm layer). This interpretation Larval behaviour and experimental approach agrees with the findings of other authors on elevated nitrification potentials in animal burrows in the subsur- Most chironomid larvae in our microcosms roamed face layers of the sediment (Kristensen et al. 1985, and ventilated the subsurface zone of the sediments at Mayer et al. 1995). Denitrifying bacteria could likewise random places. Only a few of them constructed perma- – have profited from NO3 having been introduced into nent burrows. Accordingly, obvious burrow crossings deeper sediment layers (e.g. into the 6 to 10 mm layer). by microsensors were quite rare (3 out of 81 profiles). Gilbert et al. (1998) reported a similar stimulation pat- Similar observations on the burrowing behaviour were tern of the denitrification activity by a natural commu- made at the field site from which the larvae had been nity of burrowing macrobenthos. The INT incubation collected: At a larval abundance of several 10s of 1000s technique constitutes another unspecific bulk parame- of individuals per m2, individual burrows were not ter that integrates different functional groups of distinguishable. At a lower larval abundance (100 ind. microorganisms (e.g. bacteria involved in the benthic m–2), however, almost every single larva did construct cycling of nitrogen, sulphur, iron etc.). Thus, a direct its own U-shaped burrow. One explanation for the lack correlation of the community respiration and more of spatially separated burrows could be competition for – + specific solute conversions (e.g. O2, NO3 and NH4 ) limited space in the top layer of sediments. Moreover, are not possible. the experimental sediment homogenisation had buried Aside from the enhanced nutrient supply, grazing on fresh deposits of high nutritional value into deeper sedimentary bacteria may have caused increased sediment layers, which could be exploited by sub- microbial activity. Reichelt (1991) noted that bacterivo- surface grazing instead of deposit-feeding. 184 Aquat Microb Ecol 27: 175–185, 2002

Larval burrows may also go undetected by microsen- mass, microbial community respiration and nitrogen sors because of their small inner diameter (ca. 2 mm), conversion rates. the relatively large theoretical distance between them The examination of the benthic microbial community (10 mm) and the microscale range of the microsensors. needs further refinements to address some of the

Archer & Devol (1992) postulated that O2 microdistrib- microorganisms involved in the benthic nitrogen cycle ution is governed by vertical diffusion (and not irriga- with respect to abundance and activity: The sediments tion inside burrows or radial diffusion around burrows) could be screened specifically for nitrifiers and denitri- as long as the horizontal distance between burrows is fiers using molecular techniques and inhibitors could greater than the oxic zone surrounding them. This pre- be used to specifically quantify nitrification and deni- condition was fulfilled in our case, because O2 penetra- trification. tion into the sediment surface was 2 and 5 mm (in organic-rich and -poor sediment, respectively) and O 2 Acknowledgements. Our thanks are due to Prof. Dr. D. Neu- penetration into the burrow walls can be assumed to mann, University of Cologne, Germany, who initiated and be only 40 to 70% of these values (Fenchel 1996). supervised this study. O2 microsensors were constructed by A. Thus, the presence of a few permanent larval burrows Eggers, G. Eickert and V. Hübner from Max Planck Institute was of minor importance for the shape of most vertical for Marine Microbiology, Bremen, Germany. T. Etterer, TUM Wassergüte- und Abfallwirtschaft, kindly provided the NO – concentration profiles that were interpreted using a 1D 3 ionophore. Financial support of this research project was pro- approach. The local uniqueness around individual bur- vided by the German Science Foundation with a grant to Prof. rows has been averaged out by taking microsensor Dr. D. Neumann (Ne 72/29). measurements at random spots within 1 sediment core and by slicing subcores, which were as wide as 25 mm. LITERATURE CITED Thus, the impact of Chironomus riparius on the distribution and activity of bacteria was described as laterally Alongi DM (1985) Microbes, meiofauna and bacterial produc- integrated alterations of the vertical gradients. This tivity in tubes constructed by the polychaete Capitella approach is similar to the extraction of pore-water from capitata. Mar Ecol Prog Ser 23:207–208 Archer D, Devol A (1992) Benthic oxygen fluxes on the Wash- different sediment layers that also smoothes out some ington shelf and slope: a comparison of in situ microelec- of the vertical and lateral heterogeneities around ani- trode and chamber flux measurements. Limnol Oceanogr mal burrows (Tuominen et al. 1999, Mermillod-Blondin 37:614–629 et al. 2000). The major advantage of microsensors is Bairlein F (1989) The respiration of Chironomus-larvae their superior spatial resolution in the vertical dimen- (Diptera) from deep and shallow waters under environmental hypoxia and at different temperatures. Arch sion which helps with interpreting possible inter- Hydrobiol 115:523–536 relations between larval behaviour and the benthic Blenkinsopp SA, Lock MA (1990) The measurement of elec- microbial community. In the future ion-selective micro- tron transport system activity in river biofilms. Plant Res – – + 24:441–445 sensors (NO3 , NO2 , NH4 ) should also be used inside Bowden WB (1986) Nitrification, nitrate reduction, and nitro- individual animal burrows to get some more insights gen immobilization in a tidal freshwater marsh sediment. into the microbial nitrogen conversions occurring in Ecology 67:88–99 them. So far, this approach has been performed only de Beer D, van den Heuvel JC (1988) Response of ammonium-selective microelectrodes based on the neutral with respect to O2 distribution around burrows and O2 fluctuations within burrows (Frenzel 1990, Fenchel carrier nonactin. Talanta 35:728–730 de Beer D, Sweerts JPRA, van den Heuvel JC (1991) Micro- 1996, Wang et al. 2001). electrode measurement of ammonium profiles in freshwater sediments. FEMS Microbiol Ecol 86:1–6 Fenchel T (1996) Worm burrows and oxic microniches in CONCLUSIONS marine sediments. 1. Spatial and temporal scales. Mar Biol 127:289–295 Francois F, Poggiale JC, Durbec JP, Stora G (1997) A new Our data suggest a correlation of the depth-specific approach for the modelling of sediment reworking behaviour of the chironomid larvae with changes in induced by a macrobenthic community. Acta Biotheor 45: the distribution and activity of sediment microorgan- 295–319 isms: (1) larval grazing in the deposit-feeding layer Frenzel P (1990) The influence of chironomid larvae on sediment oxygen microprofiles. Arch Hydrobiol 119:427–437 reduced the size of the microbial community as a Fukuhara H, Sakamoto M (1987) Enhancement of inorganic whole (and maybe that of nitrifiers in particular); and nitrogen and phosphate release from lake sediment by (2) larval pore-water flushing in deeper sediment lay- tubificid worms and chironomid larvae. Oikos 48:312–320 ers stimulated the growth and activity of subsurface Gilbert F, Stora G, Bonin P (1998) Influence of bioturbation on denitrification activity in Mediterranean coastal sedi- microorganisms (e.g. nitrifiers and denitrifiers) via an ments: an in situ experimental approach. Mar Ecol Prog – increased supply of O2 and NO3 . Chironomid bio- Ser 163:99–107 turbation thus produced vertical shifts of microbial bio- Hellmann H (1992) Load trends of selected chemical para- Stief & de Beer: Chironomus riparius and the benthic N-cycle 185

meters of water quality and of trace substances in the Rausch T (1981) The estimation of micro-algal protein content River Rhine between 1955 and 1988. Water Sci Technol and its meaning to the evaluation of algal biomass. I. Com- 29:69–75 parison of methods for extracting protein. Hydrobiologia Howard-Williams C, Downes MT (1993) Nitrogen cycling in 78:237–251 wetlands. In: Burt TP, Heathwaite AL, Trudgill ST (eds) Reichardt W (1988) Impact of bioturbation by Arenicola Nitrate: processes, patterns and management. John Wiley marina on microbiological parameters in intertidal sedi- and Sons, Chichester, p 141–167 ments. Mar Ecol Prog Ser 44:149–158 Johnson RK, Boström B, van de Bund WJ (1989) Interactions Reichelt AC (1991) Environmental effects of meiofaunal bur- between Chironomus plumosus L. and the microbial com- rowing. In: Meadows PS, Meadows A (eds) The environ- munity in surficial sediments of a shallow eutrophic lake. mental impact of burrowing animals and animal burrows. Limnol Oceanogr 34:992–1003 The Zoological Society of London, Clarendon Press, Karl DM, Craven DB (1980) Effect of alkaline phosphatase Oxford, p 33–52 activity on nucleotide measurements in aquatic microbial Relexans JC (1996) Measurement of the respiratory electron communities. Appl Environ Microbiol 40:549–560 transport system (ETS) activity in marine sediments: state- Karl DM, LaRock PA (1975) Adenosine triphosphate measure- of-the-art and interpretation. I. Methodology and review ments in soil and marine sediments. J Fish Res Board Can of the literature data. Mar Ecol Prog Ser 136:277–287 32:599–607 Revsbech NP (1989) An oxygen microsensor with a guard Kristensen E, Jensen MH, Andersen TK (1985) The impact of cathode. Limnol Oceanogr 34:474–478 polychaete (Nereis virens Sars) burrows on nitrification Stief P, Neumann D (1998) Nitrite formation in sediment cores and nitrate reduction in estuarine sediments. J Exp Mar from nitrate-enriched running waters. Arch Hydrobiol Biol Ecol 85:75–91 142:153–169 Lavrentyev PJ, Gardner WS, Yang L (2000) Effects of the Svensson JM (1997) Influence of Chironomus plumosus L. on zebra mussel on nitrogen dynamics and the microbial ammonium flux and denitrification (measured by the community at the sediment-water interface. Aquat Microb acetylene blockage- and the isotope pairing-technique) in Ecol 21:187–194 eutrophic lake sediment. Hydrobiologia 346:157–168 Mayer MS, Schaffner L, Kemp WM (1995) Nitrification poten- Svensson JM (1998) Emission of N2O, nitrification and denitri- tials of benthic macrofaunal tubes and burrow walls: fication in a eutrophic lake sediment bioturbated by + effects of sediment NH4 and animal irrigation behaviour. Chironomus plumosus. Aquat Microb Ecol 14:289–299 Mar Ecol Prog Ser 121:157–169 Svensson JM, Leonardson L (1996) Effects of bioturbation by Mermillod-Blondin F, Creuzé des Châtelliers M, Gérino M, tube-dwelling chironomid larvae on oxygen uptake and Gaudet JP (2000) Testing the effect of Limnodrilus sp. denitrification in eutrophic lake sediments. Freshw Biol (Oligochaeta, Tubificidae) on organic matter and nutrient 35:289–300 processing in the hyporheic zone: a microcosm method. Sweerts JPRA, de Beer D (1989) Microelectrode measure- Arch Hydrobiol 149:467–487 ments of nitrate gradients in the littoral and profundal Pelegri SP, Blackburn TH (1994) Bioturbation effects of the sediments of a meso-eutrophic lake (Lake Vechten, The amphipod Corophium volutator on microbial nitrogen Netherlands). Appl Environ Microbiol 55:754–757 transformations in marine sediments. Mar Biol 121: Tuominen L, Malela K, Kuparinen J (1999) Nutrient fluxes, 253–258 porewater profiles and denitrification in sediment influ- Pelegri SP, Blackburn TH (1995) Effects of Tubifex tubifex enced by algal sedimentation and bioturbation by Mono- (Oligochaeta: Tubificidae) on N-mineralization in fresh- poreia affinis. Estuar Coast Shelf Sci 49:83–97 water sediments, measured with 15N isotopes. Aquat van de Bund WJ, Goedkoop W, Johnson RK (1994) Effects of Microb Ecol 9:289–294 deposit-feeder activity on bacterial production and abun- Pelegri SP, Nielsen LP, Blackburn TH (1994) Denitrification in dance in profundal lake sediment. J North Am Benth Soc estuarine sediment stimulated by the irrigation activity of 13:532–539 the amphipod Corophium volutator. Mar Ecol Prog Ser Wang F, Tessier A, Hare L (2001) Oxygen measurements in 105:285–290 the burrows of freshwater insects. Freshw Biol 46:317–327

Editorial responsibility: Kevin Carman, Submitted: July 3, 2001; Accepted: January 4, 2002 Baton Rouge, Louisiana, USA Proofs received from author(s): February 15, 2002