Microbial community-level toxicity testing of linear alkylbenzene sulfonates in aquatic microcosms

FEMS Microbiology Ecology 49 (2004) 229–241 www.fems-microbiology.org

Microbial community-level toxicity testing of linear alkylbenzene sulfonates in aquatic microcosms Kristian K. Brandt

a,*

, Niels O.G. Jørgensen a, Tommy H. Nielsen a, Anne Winding

b

a

b

Section of Genetics and Microbiology, Department of Ecology, Royal Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark Department of Environmental Chemistry and Microbiology, National Environmental Research Institute, Frederiksborgvej 399, PO Box 358, DK-4000 Roskilde, Denmark Received 9 September 2003; received in revised form 21 November 2003; accepted 9 March 2004 First published online 12 April 2004

Abstract Complex microbial communities may serve as ideal and ecologically relevant toxicity indicators. We here report an assessment of frequently used methods in microbial ecology for their feasibility to detect toxic effects of the environmentally important surfactant linear alkylbenzene sulfonate (LAS) on microbial communities in lake water and treated waste water. The two microbial communities were evaluated for changes in community structure and function over a period of 7 weeks in replicated aquatic microcosms amended with various levels of LAS (0, 0.1, 1, 10 or 100 mg l
1 ) and inorganic nutrients. In general, the two communities behaved similarly when challenged with LAS. Following lag periods of 1–3 weeks, LAS was degraded to non-toxic substances. Denaturing gradient gel electrophoresis of 16S rRNA gene fragments and [3 H]leucine incorporation were the most sensitive assays with effect levels of 0–1 and 1–10 mg LAS l
1 , respectively. Community-level physiological profiles and pollution-induced community tolerance determinations using Biolog microplates demonstrated less sensitivity with effect levels of 10–100 mg LAS l
1 . Total cell counts and net uptake of inorganic N and P were unaffected even at 100 mg LAS l
1 . Interestingly, different microbial communities developed in some replicate microcosms, indicating the importance of stochastic events for community succession. We conclude that microbial community-level toxicity testing holds great promise and suggest a polyphasic approach involving a range of independent methods targeting both the structure and function of the tested microbial communities. Ó 2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. Keywords: Aquatic microbial communities; PCR-DGGE; Ecotoxicology; Linear alkylbenzene sulfonate; Microcosm

1. Introduction Bacterial communities constitute the base of aquatic food webs and are responsible for the bulk of organic matter transformations and mineral recycling in aquatic environments. Consequently, proper environmental risk assessment of toxic pollutants in aquatic environments should include studies of bacterial community responses. Such studies should aim to integrate both the fate and ecological effects of the pollutants under study. Unfortunately, this is seldom the case, as microbial toxicity assessments of pollutants released to aquatic *

Corresponding author. Tel.: +45-3528-2612; fax: +45-3528-2606. E-mail address: [email protected] (K.K. Brandt).

habitats are often based primarily on extrapolation of data from single-species, acute toxicity tests or on a few simple microbial activity measures [1,2]. In this extrapolation approach more or less arbitrarily chosen assessment factors or species sensitivity distributions (if data are available for several species) are used to calculate a predicted no-effect concentration for the entire ecosystem. Given the various sensitivities of different microorganisms to toxic chemicals and the dominance of highly diverse microbes in natural environments [3], such extrapolation approaches are likely to give erroneous results. Consequently, there is a growing interest in microbial toxicity testing at the community or ecosystem level [4–6]. Microcosm experiments with natural

0168-6496/$22.00 Ó 2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.femsec.2004.03.006

230

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

assemblages of microorganisms should provide an attractive alternative to single-species microbial toxicity tests. Relative to ecosystem-level toxicity testing, the microbial microcosm approach has several advantages such as low cost, simplicity, reproducibility, and more straight forward data interpretation. Microbial community-level toxicity testing is still not routinely performed, and there is consequently a need for baseline studies demonstrating the usefulness of this approach for ecotoxicity testing and subsequent ecological risk assessment. Ideally, community-level evaluation of pollutant-induced effects in microbial microcosms should include measures of the size, structure, and function of the microbial community under study [4]. Modern techniques targeting community structure (e.g., by DNA fingerprinting approaches) may be more sensitive compared to the more routinely used ecotoxicity assays targeting the function (e.g., activity and growth) of microbial communities. At the same time, functional redundancy is believed to be a characteristic of most microbial communities, meaning that tolerant microbial strains may replace sensitive strains with no or little loss of community functionality [5,7]. In order to compare studies of microbial community responses following pollutant exposure with the existing risk assessment literature, relevant well-studied and environmentally important model pollutants should be selected. Linear alkylbenzene sulfonates (LAS) constitute one of the best studied groups of toxic chemicals in the aquatic environment to date [2]. LAS further constitute the quantitatively most important group of xenobiotic surfactants with an annual consumption (in 1997) of approximately 2.1 million metric tons worldwide [8]. Typical LAS concentrations in untreated waste water (sewage) range between 2 and 20 mg l
1 , whereas concentrations typically are significantly below 1 mg l
1 in treated waste water due to LAS biodegradation and sorption to sludge particles in the sewage treatment plants [9,10]. Discharge of LAS-contaminated waste water leads to contamination of natural aquatic environments. Although LAS concentrations remain low in most aquatic environments, concentrations above 500 lg l
1 have been reported for Spanish estuaries receiving untreated urban waste water [11]. Locally higher concentrations may be expected, e.g., in lakes and rivers close to outlets for untreated waste water. We here report an assessment of different methods used in microbial ecology for their feasibility to detect potential adverse effects of LAS on aquatic microbial communities present in lake water and treated waste water from a municipal sewage treatment plant. The two environments were expected to differ with respect to previous exposure to LAS. Effects on microbial community structure were evaluated by denaturing gradient gel electrophoresis (DGGE) of the bacterial 16S rRNA gene amplified by PCR (PCR-DGGE), Biolog Eco-

PlateTM community-level physiological profiling (Biolog-CLPP), and by a Biolog MT PlateTM assay for pollution-induced community tolerance (Biolog-PICT). Effects on microbial community function (microbial activity and growth) were evaluated by measurement of LAS biodegradation (HPLC), dynamics of inorganic nutrients (N and P), and total bacterial cell production (incorporation of [3 H]leucine). The diversity and activity studies were supplemented by total microscopic cell counts. For comparison, we also included a single-species, LAS-sensitive toxicity assay based on a bioluminescent strain of Nitrosomonas europaea.

2. Materials and methods 2.1. Sampling and experimental set-up Lake water (LW) and treated waste water (TWW) were collected on October 22, 2000, in the southern part of the mesotrophic Lake Furesø and at the outlet of Hjortekær sewage treatment plant. Both locations are situated about 12 km northwest of Copenhagen, Denmark. Historically, Lake Furesø has received large amounts of untreated waste water, but since 1975 the pollution has been drastically reduced. Today, the lake, which covers 9.4 km2 with a mean depth of 13.5 m, only receives treated waste water from one municipal sewage treatment plant located in the northern part of the lake and rain-diluted waste water from diffuse sources. At the time of sampling, water temperatures were 14 °C (LW) and 13 °C (TWW). The water samples were transported to the laboratory in acid-cleaned polyethylene carboys (Nalge Nunc International, Hereford, UK) and kept in the dark at 15 °C until filtration the following day. In order to increase the responsiveness of the tested communities, bacterial microcosms were prepared by mixing filtered water (0.7 lm pore size GF/F filter; Whatman International, Maidstone, UK) and sterilefiltered water (0.2 lm pore size; Whatman Polycap filter capsules (including a glass fiber prefilter)) from each location in a 1:9 ratio. The bacterial microcosms were divided into volumes of 200 ml in brown 300-ml glass bottles. The bottles were divided into four triplicate series: (1) controls (no additions), (2) LAS series with addition of LAS (0.1, 1.0, 10 or 100 mg l
1 designated LAS0:1 , LAS1 , LAS10 , and LAS100 , respectively), (3) LAS + NP series with addition of LAS (0.1, 1, 10 or 100 mg l
1 ) and equivalent amounts of N (NH4 Cl) and P (Na2 HPO4 ) to achieve a C:N:P ratio of 35:7:1, and (4) NP series with addition of N and P only, at concentrations equivalent to the LAS + NP series. The following nomenclature was used for samples amended with N and P only: NP0:1 (0.2 lM N and 0.03 lM P), NP1 (2 lM N and 0.3 lM P), NP10 (20 lM N and 3 lM P), and NP100 (200 lM N and 30 lM P). Samples amended with

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

both LAS and N and P were designated LAS0:1 + NP0:1 , LAS1 + NP1 , LAS10 + NP10 , and LAS100 + NP100 . LAS were supplied from Condea Augusta (now Sasol, Milan, Italy) as a 25.5% (w/w) solution (Isorchem 113) consisting of C10 –C13 homologues with an average chain length of 11.6. The microcosms were incubated on a rotary shaker in the dark at 15 °C. One hour after the beginning of the incubations and at various time points during the incubation period (up to 43 days), subsamples were taken for measurement of chemical and microbiological parameters. Samples for chemical analysis (i.e., measurements of LAS and inorganic N and P), N. europaea toxicity bioassay, leucine incorporation assay, and direct cell counts were taken at regular intervals (see figures) during the microcosm incubations, until all or most LAS had disappeared. Community analysis by PCR-DGGE, Biolog-CLPP, and Biolog-PICT were performed only on the LW and TWW samples used for setting up the microcosms (Day 0) and once after 24 or 43 days of incubation when complete LAS removal had occurred. Hence, microcosms with 0.1 or 1 mg LAS l
1 were sampled after 24 days, while microcosms with 10 mg l
1 were sampled after 43 days. Microcosms with 100 mg LAS l
1 were also sampled after 43 days (only PCRDGGE and Biolog-CLPP), despite the fact that no LAS removal had occurred. A few microcosms were sampled after both 24 and 43 days to evaluate the effect of time on community succession. 2.2. Chemical analysis Subsamples of 1 ml for analysis of LAS were filtered through 0.2-lm Minisart membrane filters (Sartorius AG, G€ ottingen, Germany). The samples were added to 1 ml 100% methanol (HPLC grade) to stop further microbial activity and reduce LAS adhesion to the walls of the polypropylene vials before freezing at )20 °C. LAS homologues (C10 –C13 ) were analyzed by HPLC using the Hewlett–Packard Model 1100 and a LiChroCART 250-4 HPLC-Cartridge with LiChrosphere 100 RP-18 (5-lm) particles (Merck Eurolab, Darmstadt, Germany) at 40 °C. Solvents for the analysis consisted of the following solutions: aqueous NaClO4 (14 g l
1 ), aqueous trifluoroacetic acid (140 mg l
1 ), acetonitrile (HPLC grade), and Milli-Q water. The solvents were mixed in a multi-step gradient protocol modified from [12]. The initial composition was 5%, 28%, 40%, and 27% NaClO4 , TFA, acetonitrile, and Milli-Q, respectively, followed by 20%, 0%, 60%, 20% at 5 min, 0%, 0%, 90%, 10% at 15 min, 0%, 17%, 60%, 23% at 18 min, and 5%, 28%, 40% and 27% at 20 min. The flow rate was 1 ml min
1 . LAS was detected using a HP diode array detector operated at 225 nm and in line with a HP 1090 fluorescence detector at 225 nm excitation and 295 nm emission wavelengths. The different LAS homologues in

231

the commercial LAS mixture and in the microcosm samples were integrated before calculating a total LAS concentration. The detection limit of the LAS mixture used in the experiment was approximately 0.1 mg LAS l
1 (fluorescence detection). Subsamples of 1 ml for analysis of inorganic N and P were filtered through 0.2-lm Minisart membrane filters (Sartorius AG, G€ ottingen, Germany). NHþ 4 , NO2 + 2þ
NO3 , and PO4 were measured on an AlpKem FlowSolution IV autoanalyzer (OI Analytical, College Station, TX) according to procedures recommended by the manufacturer. 2.3. Nitrosomonas europaea bioluminescence toxicity assay A bioluminescent strain of the autotrophic ammoniaoxidizing bacterium N. europaea was used for testing the toxicity of water samples from the experimental microcosms. The used N. europaea ATCC 19718 (pHLUX20) reporter strain constitutively expresses luxAB genes from Vibrio harveyi and the resulting bioluminescence is positively correlated to the metabolic activity (i.e., ammonia oxidation rate) of the cells [13]. Microcosm water samples (750 ll) were collected at the same days as for HPLC analysis of LAS and stored at )18 °C in transparent 2-ml polypropylene microtubes (AxyGen, Inc., Union City, CA). The day before the actual performance of the toxicity assay, the samples were thawed for 10 min in a temperature block at 60 °C and subsequently frozen again. On the following day, the thawing step was repeated in order to kill or inactivate possible bacterial grazers in the water samples. N. europaea cells were cultivated in 250-ml batches in autotrophic growth medium [14,15]. An early stationary phase N. europaea test culture was centrifuged (13,000g, 20 min, 10 °C) and resuspended in 20% of the original volume of fresh ammonium-free medium. Incubations were initiated by adding 200 ll of the concentrated test cell suspension and 15 ll of 0.5 M (NH4 )2 SO4 to each microcosm sample, which were incubated on a rotary shaker (300 rpm, 6 h, 22 °C). After the incubation, bioluminescence was measured by luminometry (BioOrbit 1253, Merck Eurolab) as described previously [15]. 2.4. Leucine incorporation rates Bacterial production was measured in 1-ml triplicate samples that received a mixture of [3 H]leucine and unlabeled leucine to a final concentration of 100 nM [16,17]. Killed controls containing 2% formaldehyde (final concn.) were also added to [3 H]leucine and served as blanks. After incubation periods of 30 min at 15 °C, the samples were mixed with 120 ll 50% trichloroacetic acid (TCA), centrifuged, and re-extracted in 5% TCA [18,19]. Finally, the precipitates were redissolved in 1 ml

232

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

of a scintillation cocktail (Lumasafe, Lumac, Groningen, The Netherlands) and assayed for radioactivity. Bacterial assimilation rates for carbon and nitrogen were derived from leucine incorporation data using previously reported conversion factors [20] assuming a bacterial C:N ratio of 5 [21] and a cell volume of 0.1 lm3 . 2.5. Cell enumeration and dominant cell morphologies For determination of bacterial abundance and dominant cell morphologies, 1-ml samples were collected from the microcosms and preserved with formaldehyde to a final concentration of 2%. The bacteria were observed by epifluorescence microscopy (Zeiss Axioplan equipped with a 100 Plan-Neofluar objective and an HBO 100 W lamp) after staining with acridine orange [22]. At least 200 cells were counted in each of 10 different fields on the filters. Photomicrographs of predominant cell morphologies were recorded using a Nikon CoolPix 990 3.2 Mbit digital camera, mounted on the microscope. 2.6. Denaturing gradient gel electrophoresis For PCR-DGGE analysis of the bacterial 16S rRNA gene, 15 ml of water sample was filtered through Gelman Supor 200 filters (0.2-lm pore size; Fisher Scientific Int., www.fisherscientific.com). The filters were rinsed three times with 3 ml TE buffer (10 mM Tris in 1 mM EDTA, pH 7.5) before storage, initially at )20 °C and later at )80 °C until extraction. Nucleic acids were extracted from one quarter of each filter, cut into pieces, and vortexed in 100 ll TE buffer at pH 8.0. However, bacterial DNA in LW and TWW at the start of the experiment was extracted from entire filters in 200 ll TE buffer. Bacteria in the suspension were lysed by the following freeze–thaw–boiling protocol: 5 min at )80 °C, 10 min at 20 °C, and 10 min at 102 °C in a heating block. Cell debris and filter pieces were removed by centrifugation (15,000g for 10 min). Finally, DNA was purified by the High Pure PCR Template Preparation Kit (Roche Cat. no. 1 796 828, Roche A/S, Hvidovre, Denmark) and stored at 4 °C before PCR. A fragment of bacterial 16S rRNA gene (pos. 968–1378 by Escherichia coli numbering) was PCR-amplified [23]. The yield of the PCR and the size of the products were analyzed on a 1% agarose gel followed by fluorometric PCRproduct quantification by PicoGreen (Molecular Probes, Invitrogen A/S, T
astrup, Denmark). Fifty nanograms of each PCR product was loaded in separate lanes on 40–60% denaturing gradient gels and electrophoresed, silver stained, and digitized as described previously [23]. The number of bands and their position on the DGGE gel were analyzed and dendrograms were created based on Dice coefficient of similarity and the

UPGMA method by DendronÒ (Solltech Inc., Oakdake, IA). 2.7. Community level physiological profiling Biolog EcoPlateTM microplates (Biolog, Inc., Hayward, CA) were used to determine differences in the metabolic potential of bacterioplankton in the different microcosms. The EcoPlates contain three replicate wells of 31 different carbon sources [24] and three replicate control wells without any carbon source. The plates were inoculated with 150 ll undiluted culture sample. Three replicates from each different replicate microcosm were used per treatment. The plates were incubated at 15 °C on a rotary shaker (200 rpm) in the dark. The optical density (c ¼ 590 nm) of each well was measured immediately and subsequently every 24 h for 7 days using a multi-well plate reader (EL312e, Bio-Tek Instruments, Winooski, VT). Analysis of plate data was done by the average well color development (AWCD) set-point method as recommended by Garland and coworkers [25–27]. Initially, the color development in each well was corrected by subtracting the absorbance of the corresponding control well. Plate readings with an AWCD of 0.35  0.05 were subsequently chosen for statistical data analysis after prior normalization of the data by dividing the corrected absorbance values with the corresponding AWCD. Negative responses were set to zero prior to ordination. The number of positive wells were followed over time and defined as wells with an optical density (after subtracting control values) of 0.25 or higher. Principle component analysis (PCA) was performed on the normalized data using the Systat 10 software package (SPSS, Inc., Chicago, IL). 2.8. Pollution-induced community tolerance The determination of pollution-induced community tolerance (PICT) has been suggested as a sensitive and ecological relevant measure of adverse effects of toxicants exerting a selection pressure on microbial communities [5]. In this study, we took advantage of the BiologTM test system to investigate evolution of PICT (Biolog-PICT) [28–30]. Briefly, the aquatic microbial community of each microcosm (127.5 ll sample per well) was inoculated in Biolog MTTM microplates (i.e., microplates with no preadded growth substrates, but with the same patented nutrient base and redox dye chemistry as other Biolog microplates; Biolog, Inc., Hayward, CA). Inoculated cells were exposed to 3  8 wells containing glucose (800 mg l
1 in each well) and eight different LAS concentrations (0, 0.3, 1, 3, 10, 30, 100 or 300 mg l
1 ), implying that three replicate LAS dose–responses could be obtained per microcosm. In addition, the aquatic microbial community of each microcosm sample was exposed to 1  8 wells containing no glucose and the same eight different LAS levels.

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

2.9. Statistics

Lake water

Treated waste water

1.5 LAS1

A

LAS1

C

LAS1 + NP1

LAS1 + NP1

1.0

0.5 -1

LAS (mg l )

The latter incubations served as negative controls and verified that the different LAS levels did not interfere with color development in the wells. The Biolog MT microplates were incubated and read as described above for the Biolog EcoPlates, except that each plate was read at least twice every 24 h. Analysis of plate data was performed according to Rutgers et al. [28]. Hence, the microbial activity/growth in each microplate well was determined as the maximal rate of color development observed within a 24-h interval during the one week-long incubations.

233

0.0 LAS10

B

LAS10

D

LAS10 + NP10

LAS10 + NP10

10

The effect of experimental treatments (LAS and NP levels) on leucine incorporation rates was tested using a two-way analysis of variance (ANOVA) using SigmaStat Version 3.0 (SPSS, Inc., Chicago, IL). The Tukey test was used for all pairwise multiple comparisons of the mean responses to the different comparable experimental treatment groups (e.g., control, NP1 , LAS1 , and LAS1 + NP1 ). When required, the raw data were transformed as suggested by SigmaStat output. Dunnett’s test was used for comparison of LAS treatments with the corresponding control to estimate no-observed-effect-concentrations (NOEC) and lowest-observed-effectconcentrations (LOEC). Biolog-PICT data were analyzed by non-linear regression analysis to estimate 20% inhibition effect concentrations (EC20 ) and corresponding 95% confidence intervals. ToxcalcTM Version 5.0 (Tidepool Scientific Software, McKinleyville, CA) was used for estimation of LOEC, NOEC, and EC values.

3. Results 3.1. LAS biodegradation LAS degradation occurred in all LAS amended microcosms (Fig. 1), except in microcosms receiving 100 mg LAS l
1 , where LAS concentrations remained unchanged (data not shown). LAS degradation was initiated after a lag period lasting from about 7 days (1 mg l
1 treatments) and up to 14–21 days (10 mg l
1 treatments). After onset of the degradation, an almost complete LAS removal occurred in almost all microcosms. There was no significant effect of enrichment with N and P on the degradation kinetics, suggesting that the microbial LAS degraders were not limited by N and P. This observation contrasts with data from an earlier pilot experiment using a similar set-up and water collected at Lake Furesø just two months prior to our sampling. In this experiment, LAS degradation was significantly stimulated by N and P addition (N.O.G. Jørgensen et al., unpublished results). Importantly, these data show that the functional response of a microbial community may change depending on the time of sampling.

Sickle-shaped cells

5 Short rod-shaped cells

0 0

10

20

30

40

0

10

20

30

40

50

Incubation period (days) Fig. 1. LAS degradation curves in individual microcosms measured by HPLC. (A) LW amended with 1 mg LAS l
1 ; (B) LW amended with 10 mg LAS l
1 ; (C) TWW amended with 1 mg LAS l
1 ; (D) TWW amended with 10 mg LAS l
1 . Microcosms (n ¼ 3) were amended with LAS only (closed symbols) or LAS + NP (open symbols). In (D), individual TWW microcosms dominated by short rod-shaped and large sickle-shaped cells are indicated (see text).

3.2. Nitrosomonas europaea toxicity assay The initial LAS level in each microcosm strongly affected the bioluminescence output from the N. europaea test cells, as bioluminescence was progressively inhibited at higher LAS concentrations. In the 100 mg LAS l
1 microcosms, a constant and almost complete (>90%) inhibition of light output was observed (data not shown) reflecting the absence of LAS degradation in these treatments. In microcosms amended with 10 mg LAS l
1 , the toxicity decreased over time with no residual toxicity remaining after 43 days. The toxicity reduction in each individual microcosm was tightly correlated to LAS removal as measured by HPLC, indicating that LAS was degraded to non-toxic substances. No apparent inhibition was observed when testing samples from the 0.1 and 1 mg LAS l
1 microcosms (data not shown). Similarly, no inhibition of bioluminescence was observed in samples taken from LAS-free microcosms with different concentrations of inorganic nutrients (N and P; data not shown). 3.3. Cell production rates as determined by leucine incorporation Cell production rates as estimated from [3 H]leucine incorporation rates varied in a highly dynamic manner,

234

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

but showed rather similar patterns for both LW and TWW microcosms (Fig. 2). Immediately after set-up of the microcosms, the bacterial production was <0.1  106 cells l
1 h
1 . After 3 days, the cell production rate had increased by 50- (TWW) to 100-fold (LW) in all microcosms, except those amended with 100 mg LAS l
1 (both LW and TWW) or 10 mg LAS l
1 (only TWW), where a much weaker growth stimulation was observed (Fig. 2). During the following days cell production rates decreased, again with the exception of all the 100 mg LAS l
1 microcosms and TWW microcosms containing 10 mg LAS l
1 , in which leucine incorporation rates tended to be more constant (albeit with significant fluctuations) or even increased over time. A toxic effect of high LAS levels (10 and 100 mg l
1 ) was evident in both LW and TWW microcosms with significantly decreased microbial growth rates during the first 7–14 days in an LAS concentration-dependent manner (Fig. 2B, C, E and F; p < 0:01). By contrast, microbial growth was stimulated (p < 0:05) after 29 or 36 days in LW and TWW microcosms amended with 10 mg LAS l
1 , probably because LAS or its degradation

Lake water 8

A

-1 6 -1

Bacterial production (cells × 10 l h )

6 4 2

Treated waste water

Control NP1

5

Control NP1

D

4

LAS1

LAS1

LAS1 + NP1

LAS1 + NP1

3 2 1

0

Control NP10

B 6 4

0

Control NP10

E

4

LAS10

LAS10

LAS10+NP10

LAS10+NP10

3 2

2

1

0

Control NP100

C 6 4

Control NP100

F

LAS100

LAS100

LAS100+NP100

LAS100+NP100

0 4 3 2

2

1

0

intermediates served as growth substrates. A transient stimulation of microbial growth only at Day 29 was observed for three out of four experimental treatments with 100 mg LAS l
1 (Fig. 2C and F). Addition of N and P to microcosms did not increase leucine incorporation rates (Fig. 2), indicating that N and P were not limiting for bacterial growth. Net dynamics of inorganic nutrients (i.e., NHþ 4,
3
NO
3 + NO2 , or PO4 ) remained unaffected by LAS in both LW and TWW microcosms (data not shown). NHþ was initially present at a concentration of 4 2.2 7lM in the LW microcosms without added nutrients. Concomitant with the high bacterial cell production during the first week (Fig. 2), all NHþ 4 was briefly depleted. Apart from this, no signs of detrimental nutrient depletions were observed in any of the microcosms. 3.4. Cell numbers and morphology The bacterial density in the microcosms was initially between 2  108 and 3  108 cells l
1 and increased to between 3  108 and 189  108 cells l
1 during the 43-day incubation period (data not shown). Large fluctuations in the densities were observed. Hence, the number of bacteria in the microcosms did not reflect the actual bacterial cell production. Interestingly, different cell morphologies were found to dominate replicate TWW microcosms amended with 10 mg LAS l
1 (Fig. 3). In the three microcosms receiving N and P and in one of the microcosms without added N and P, the microbial community was dominated by short rods, while larger and sickle-shaped cells made up >90% of the bacterial cells in the remaining two microcosms. The observed cell morphology differences between replicate microcosms were also reflected in the community function and structure. Increased microbial activity in microcosms dominated by short rods was indicated by faster microbial degradation of LAS (Fig. 1D) and by higher leucine incorporation rates at Day 43 (2.1  106 –3.1  106 cells l
1 h
1 versus only 1.2  106 –1.4  106 cells l
1 h
1 in microcosms dominated by large, sickle-shaped cells; Fig. 3). Altered microbial community structure was also indicated by DGGE band profiles (see below).

0 0

10

20

30

40

0

10

20

30

40

Incubation period (days) Fig. 2. Bacterial production rates (means  SDs) as measured by the [3 H]leucine incorporation technique. (A) LW amended with 1 mg LAS l
1  NP; (B) LW amended with 10 mg LAS l
1  NP; (C) LW amended with 100 mg LAS l
1  NP; (D) TWW amended with 1 mg LAS l
1  NP; (E) TWW amended with 10 mg LAS l
1  NP; (F) TWW amended with 100 mg LAS l
1  NP. Microcosms (n ¼ 3) were amended with LAS only (closed triangles) or LAS + NP (open triangles). Corresponding control microcosms (n ¼ 3) were not amended with LAS or NP (closed circles) or with NP only (open circles).

3.5. Bacterial community structure analysis by PCRDGGE Analysis of DGGE band profiles based on PCRamplified fragments of the 16S rRNA gene revealed structural changes of the bacterial communities. LW and TWW samples were collected at Day 0 immediately before microcosm setup, at Day 24 (1 mg LAS l
1  NP1 and appropriate controls  NP1 ) and at Day 43 (10 mg LAS l
1  NP, 100 mg LAS l
1  NP, and the

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

235

Fig. 3. Photomicrographs and DGGE profiles documenting the development of different microbial communities in replicate TWW microcosms (10 mg LAS l
1 treatment) following 43 days of incubation. (A) Photomicrograph of microcosm dominated by large sickle-shaped cells; (B) DGGE profiles of microcosm communities dominated by large sickle-shaped cells; (C) DGGE profiles of microcosm communities dominated by small rodshaped cells; (D) Photomicrograph of microcosm dominated by small rod-shaped cells.

appropriate controls NP10 or NP100 ). To evaluate the effect of time, one experimental treatment (NP10 ) was sampled after both 24 and 43 days. The number of bands in each lane of the DGGE gels was found to vary between 3 and 25 (Table 1). A single band is expected to represent a unique 16S rRNA gene sequence, although some comigration of multiple bands may occur [31]. For the TWW microcosms, the number of bands within the lanes was rather stable for all microcosm except for the microcosms receiving 100 mg LAS l
1 (Table 1) indicating a reduced genetic diversity at this high LAS concentration. The few bands present were very distinct and clear. For the LW microcosms, a moderate increase of genetic diversity was indicated in microcosms amended with low levels of LAS and/or inorganic nutrients, whereas amendment with 100 mg LAS l
1 resulted in a lower number of DGGE bands comparable to or slightly lower than the number found for the control treatment.

Similarity of the samples was calculated using cluster analysis allowing for construction of similarity dendrograms (Fig. 4). To facilitate the interpretation, gels with LW and TWW samples were analyzed separately. Similarly, samples from Day 24 and 43 were depicted in separate dendrograms. Data from Day 0 were included in both the Day 24 and Day 43 results to facilitate comparison between dendrograms. DGGE band profiles generated from replicate microcosms generally showed higher similarity to each other, than to profiles from the other experimental treatments. Exceptions were the LAS-free LW microcosms amended with N and P after 24 days (Fig. 4A), the 10 mg LAS l
1 LW microcosms after 43 days (Fig. 4B), and the 10 mg LAS l
1 TWW microcosms after 43 days (Fig. 4D). In these three experimental treatments, the obtained DGGE band profiles thus exhibited a high variability with similarity levels as low as 30–40% between replicate microcosms. Interestingly, the dendrogram branching pattern of the six

Table 1 Average number of bands in PCR-DGGE gels Treatment

Lake water 0 days

Start Control NP1 LAS1 LAS1 + NP1 NP10 LAS10 LAS10 + NP10 NP100 LAS100 LAS100 + NP100

Treated waste water 24 days

0 days

24 days

43 days

a

19

16 11 24 19 17 12

(8–14) (23–25) (16–21) (14–19) (12)

Ranges of triplicates are indicated in brackets. No replicates. b Duplicates. a

43 days

a

14 15 11 18 8 11

(11–16) (13–17) (10–13) (14–22) (6–10)b (9–13)b

16 17 18 18 17

(15–17) (17) (17–19) (18–19) (14–19)b

19 15 14 14 4 6

(18–20) (13–18) (11–16) (13–14) (3–6) (5–7)b

236

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

Fig. 4. Dendrograms showing similarity values (Dice coefficient of similarity) calculated by cluster analysis (UPGMA method) based on PCR-DGGE fingerprinting. (A) LW microcosms amended with 1 mg LAS l
1 and corresponding controls with or without NP amendment sampled after 24 days; (B) LW microcosms amended with 10 or 100 mg LAS l
1 and corresponding controls with or without NP amendment sampled after 43 days; (C) TWW microcosms amended with 1 mg LAS l
1 and corresponding controls with or without NP amendment sampled after 24 days; (D) TWW microcosms amended with 10 or 100 mg LAS l
1 and corresponding controls with or without NP amendment sampled after 43 days. LW t0 and TWW t0 represent the initial samples used for construction of LW and TWW experimental microcosms, respectively. For additional sample nomenclature, see Section 2.1.

TWW microcosms amended with 10 mg LAS l
1 ( NP) was highly consistent with microscopic observations of the dominant cell morphologies (Figs. 3 and 4D). Despite the observed replicate variability, the obtained community fingerprints convincingly demonstrated that different microbial communities developed as a result of LAS and NP additions. In the LW microcosms, low LAS treatments of 1 mg l
1 formed one distinct cluster and showed low similarity values (<40%) when compared to comparable treatments without LAS and with the LAS + NP additions separating the main cluster of LAS addition at a secondary level (Fig. 4A). Similarly, all TWW microcosms amended with 1 mg LAS l
1 clustered together, although similarity values between the LAS

treatments and the comparable treatments without LAS were much higher (>70%; Fig. 4C). However, the Day 0 control (TWW t0 ) did not cluster with any of the microcosm treatments sampled after 24 days (50% similarity; Fig. 4C). This indicates that microbial communities in TWW microcosms to some degree changed their composition in the same direction irrespective of experimental treatment. By contrast, the corresponding Day 0 LW control (LW t0 ) clustered together with all samples from microcosms not receiving LAS at Day 24 (Fig. 4A). This indicates that the LAS addition was the primary factor responsible for shaping the microbial community structure in the LW microcosms after 24 days. After 43 days, the similarities between microcosm communities

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

(both between treatments and replicates) were generally lower (Fig. 4B and D). However, LW and TWW microcosms amended with LAS  NP clustered separately and distinct from microcosms amended with N and P only. The lower similarity values between experimental treatments analyzed after 43 days could probably be explained by the more extreme nature of the experimental treatments (i.e., higher LAS and NP additions) when compared to treatments analyzed after 24 days. The longer time period available for microbial community succession most likely did not play an important role. Hence, when comparing DGGE band profiles from the same microcosms (data only available for NP10 treatment) after 24 and 43 days, no time-dependent sample clustering was observed (data not shown). By contrast, a weak time-dependant clustering of the corresponding samples taken from LW microcosms was observed (data not shown). However, the overall similarity values between experimental treatments were similar in LW microcosms sampled after 24 and 43 days. 3.6. Community level physiological profiling The functional diversity of the bacterial communities was analyzed using Biolog EcoPlatesTM and principal component analysis (PCA) on the same samples and dates as used for PCR-DGGE. The number of substrates supporting growth after 7 days of incubation in Biolog EcoPlatesTM ranged from 16.3  3.5 to 26.7  2.5 and 12.0  2.7 to 24.7  1.5 for the LW and TWW treatments, respectively, out of 31 possible. However, there were no obvious trends with regard to number of substrates utilized and effects of LAS or N and P additions (data not shown).

The microbial communities in the LW microcosms were clearly separated into two major clusters when applying PCA on the Biolog-CLPP data (Fig. 5A). Cultures with 100 mg LAS l
1  NP harbored microbial communities that clearly differed from those present in all the other cultures. This conclusion was supported by the high scores obtained for the first two principal components (PCs) accounting for 66.3% and 15.9% of the variance in the entire data set. Attempts to separate any of the remaining treatments from each other by PCA after omitting the data from cultures with 100 mg LAS l
1 were not successful (data not shown). Microbial communities in the TWW microcosms were also only partially separated from each other after PCA of CLPP data. Again high scores were obtained for the first two PCs accounting for 61.2% and 14.0% of the total variance. Only microcosms with 100 mg LAS l
1 formed a distinct cluster (Fig. 5B), whereas the remaining treatments formed a rather loose cluster. As for the LW samples, the remaining treatments could not be clearly distinguished from each other by PCA even after omitting the 100 mg LAS l
1 microcosm data prior to ordination. Hence, the replicate variability was of equal magnitude as the variability between experimental treatments. However, there appeared to be a tendency for lower PC1 scores with increasing N and P additions, whereas LAS concentrations up to 10 mg l
1 had no obvious effects on the PC scores. To some extent the ability to separate different microcosm communities by PCA of CLPP data could be related to the utilization of individual growth substrates. This was especially the case for the LW treatments. Utilization of the substrates b-methyl-D -glucoside, D galactonic acid, c-lactone, D -galacturonic acid, putres-

Treated waste water

All other samples

A

PC2 (15.9%)

0.2

0.6

NP1

B

All other samples NP100

0.0

0.4 0.2 0.0

-0.2

-0.2 -0.4

NP10

-0.6

-0.6

-0.8 -1.0 -0.8

LAS10+NP10 -0.4

LAS100±NP100

-0.8

LAS100±NP100 -0.6

-0.4

-0.2

0.0

PC1 (66.3%)

0.2

-0.4

-0.2

0.0

PC2 (14.0%)

Lake water 0.4

237

0.2

-1.0 0.4

PC1 (61.2%)

Fig. 5. Principal component score plots of Biolog-CLPP data subjected to principal component analysis. (A) LW microcosm data; (B) TWW microcosm data. Microcosms amended with 1 mg LAS l
1 and corresponding controls with or without NP amendment were sampled after 24 days, while microcosms amended with 10 or 100 mg LAS l
1 and corresponding controls with or without NP amendment were sampled after 43 days. The control microcosms were sampled after both 24 and 43 days. See Fig. 4 for sample nomenclature.

238

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

cine, L -asparagine, and D -glucosaminic acid were thus positively correlated (r2 > 0:7) to PC1, while the substrates glycogen, D ,L -a-glycerol phosphate, and N acetyl-D -glucosamine were negatively correlated (r2 > 0:7) to PC1. In case of the TWW treatments, only the substrates D -cellubiose and glucose-1-phosphate showed a strong correlation (r2 > 0:7) to PC1. The highlighted substrates showing high correlations to PC1, in turn explaining 66.3% and 61.2% of the total variance were thus the most important for differentiating between samples from different experimental treatments. 3.7. Pollution-induced community tolerance Evolution of PICT (i.e., increased community tolerance to LAS following LAS exposure) was evaluated by comparing the LAS tolerance of glucose-utilizing microorganisms sampled from different microcosm treatments. For all treatments, color formation in the Biolog MT microplates were not significantly affected even with 100 mg LAS l
1 (i.e., NOEC P100 mg l
1 ), making it impossible to obtain robust EC values, as we only analyzed one concentration (300 mg l
1 ) above this level. Hence, there was no significantly increased community tolerance following LAS exposure among BiologTM culturable, glucose-utilizing microorganisms.

4. Discussion 4.1. Fate of LAS in microcosms Although we only measured the primary LAS degradation by HPLC analysis, it follows from the N. europaea bioluminescence data that LAS at 10 mg l
1 was either mineralized or degraded to non-toxic intermediates. The increased leucine incorporation rates observed when almost all LAS was removed in the 10 mg LAS l
1 microcosms towards the end of the incubations suggest that some production of non-toxic, organic intermediates took place, allowing for a subsequent microbial growth. 4.2. LAS ecotoxicity testing in microcosms There were no drastic differences in the responses of microbial communities in TWW and LW following LAS exposure. However, a slightly more robust community in TWW was indicated by the more similar DGGE band profiles (i.e., higher similarity values) following different LAS treatments in TWW as compared to in LW microcosms (Fig. 4A and C) and by the faster recovery of bacterial production in TWW following incubation with 100 mg LAS l
1 (Fig. 2). The resistance to LAS of the studied microbial communities varied between the different measured test parameters as summarized from the estimated NOEC

and LOEC values presented in Table 2. In general, measures of community structure and function served as better toxicity indicators than the applied measure of community size. This is consistent with ecological theory predicting the replacement of sensitive species with tolerant ones before any decrease of community size or biomass [5,7] occurs. Hence, community profiling approaches may serve as promising early warning ecotoxicity indicators. Microbial in situ growth as estimated from leucine incorporation showed a significant inhibition at 10 mg LAS l
1 , whereas lower concentrations (0.1 or 1 mg l
1 ) had no significant effect (Table 2). These data are comparable to thymidine incorporation data from a related study employing in situ enclosures in a Danish lake [32], showing a toxic effect with 5 mg LAS l
1 , but not with 1 mg l
1 . However, in the latter study, evidence was found for a LAS-dependent stimulation of microbial growth at very low concentrations (0.1 or 0.25 mg l
1 ). Other researchers [33] have reported a somewhat higher LAS toxicity (EC50 of 1.21 and 1.07 mg l
1 in freshwater and sea water, respectively). By contrast, up to 3 mg LAS l
1 did not affect the [3 H]amino acid incorporation rates in an experimental stream facility [34]. Hence, it can be concluded that leucine incorporation and related methods measuring in situ growth represent sensitive assays for investigating adverse effects of LAS on the function of microbial communities. This is consistent with the notion of growth being an integrating effect parameter, which incorporates toxic effects irrespective of toxicant mode of action [5] and with experimental data showing that bacterial growth may be more sensitive to LAS exposure than bacterial energy metabolism [14]. PCR-DGGE community fingerprinting in combination with cluster analysis appeared to be the most sensitive test parameter followed by leucine incorporation and N. europaea toxicity assay (Table 2). Hence, a clear effect of 1 mg LAS l
1 on DGGE band profiles was observed, indicating that the bacterial community structure was affected by LAS. By contrast, BiologCLPP in combination with principal component analysis was rather insensitive with LOEC’s of 100 mg l
1 (Table 2). Diversity estimates based on the number of 16S rRNA gene bands in DGGE gels (genetic diversity) or number of utilized substrates in Biolog EcoPlates (functional diversity) were less sensitive to LAS as compared to the corresponding ‘‘community fingerprints’’ (i.e., cluster analysis of DGGE band profiles and PCA of Biolog substrate utilization patterns; Table 2). In conclusion, our results suggest that PCR-DGGE community fingerprinting may be a highly sensitive and useful indicator of toxicant-induced community changes. Although community composition as evaluated by PCR-DGGE fingerprinting may be very sensitive to toxicant exposure, this approach suffers from one

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

239

Table 2 Ecotoxicological test parameters estimated for lake water (LW) and treated waste water (TWW) microcosms amended with LAS only Parameter

NOECa (mg LAS l
1 )

LOECa (mg LAS l
1 )

LW

TWW

LW

TWW

>100

>100

nd

nd

Community function LAS biodegradationa Duration of lag Inhibition

1 10

1 10

10 100

10 100

Nutrient dynamics Phosphate Nitrate + Nitrite Ammonium

>100 >100 >100

>100 >100 >100

nd nd nd

nd nd nd

Leucine incorporation rate Day 3

1

1

10

10

Community structure PCR-DGGE Diversity (number of bands) Composition (cluster analysis)a

>100 0

10 0

nd 1

100 1

Biolog-CLPP Diversity (number of utilized substrates) Composition (principal component analysis)a Biolog-PICT

>100 10 >10

>100 10 >10

nd 100 nd

nd 100 nd

Single-species N. europaea toxicity assay (Day 3)

1

1

10

10

Community size Cell numbers Day 3

NOEC, highest no-observed-effect concentration; LOEC, lowest-observed-effect concentration. nd; not determined. a NOEC and LOEC estimates were based on visual examination of the obtained graphs (LAS degradation, Biolog-CLPP PCA score plots) or dendrograms (PCR-DGGE cluster analysis).

important drawback for the specific purpose of toxicity testing. Hence, a change of PCR-DGGE profile following exposure to a test chemical does not imply whether the chemical is toxic or not. Only a minor fraction of indigenous microorganisms are considered capable of degrading LAS in natural environments [35– 37], and the changed DGGE band patterns at low LAS concentrations may thus simply reflect the growth of microbial consortia responsible for biodegradation of the surfactant [36–40]. PCR-DGGE profiling and related PCR-based community fingerprinting approaches should thus primarily be considered indicators of disturbance rather than direct toxicity. PCR-DGGE fingerprinting only indicates a direct toxic effect on the microbial community in extreme cases, e.g., the reduced number of DGGE bands observed for the 100 mg LAS l
1 treatments. By contrast to PCR-DGGE and Biolog-CLPP, the PICT concept may allow for the determination of a causal linkage between toxicant exposure and altered community structure [5]. However, the applied BiologPICT approach failed to detect any significant effects of

LAS in the present study (Table 2). This does not necessarily mean that toxicant-induced selection did not occur in the microcosms. Hence, it should be emphasized that only the culturable fraction of bacteria capable of growing in Biolog MT microplates at high substrate concentrations was monitored for development of PICT. Given the observed failure of the Biolog system to detect apparent changes in microbial community structure following LAS exposure by BiologCLPP, it is thus likely that the applied Biolog-PICT approach also missed some of the dominant microbial population responses to toxicant exposure. Another potential problem of the Biolog-PICT approach relates to the rather long PICT detection phase (several days of incubation in Biolog plates), which implies the risk of enriching for both the ability to use the added substrate and to tolerate the added toxicant [5]. Consequently, future research on the PICT approach for communitylevel toxicity testing should preferably focus on shortterm (hours) incubations under more realistic test conditions than provided by the Biolog microplate system. Based on our results, the leucine incorporation

240

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241

technique seems very attractive for this purpose. It is thus a sensitive, short-term assay that can be carried out under realistic test conditions and further has the inherent advantage of being an integrating effect indicator making it suitable for toxicants with different modes of action. 4.3. Stochastic events involved in succession of microbial communities One very interesting finding of our investigation was the development of completely different microbial communities in some of the replicate microcosms (Figs. 3 and 4). Stochastic events involved in the set-up of microcosms and/or during subsequent community succession represent the most likely explanation for this phenomenon. Hence, the replicate microcosm communities may initially have been slightly different due to stochastic events involved in preparing the 200 ml microcosms from only 20 ml of pre-filtered (0.7-lm pore size) water, and subsequent community succession may have amplified these differences. Candidate key populations involved in these stochastic events include grazing protozoa and bacteria involved in LAS degradation. The initial numbers of these specific populations were expected to be low, but the numbers probably increased significantly due to the conditions conducive for their proliferation. The hypothesis of varying protozoa abundance is consistent with the direct observation of protozoa in only some of the microcosms by microscopy, although the actual numbers of protozoa were too low to perform accurate direct counts. The hypothesis of a variable abundance of LAS-degrading microorganisms is consistent with the different LAS degradation kinetics observed for replicate microcosms dominated by small rod-shaped cells and large sickle-shaped cells, respectively (Fig. 1D). 4.4. Perspectives for microbial community-level toxicity testing Community-level toxicity testing of chemicals holds great promise and should supplement single-species laboratory assays when testing potential harmful chemicals. We admit that stochastic events involved in the succession of natural microbial communities may introduce difficulties in performing controlled, replicated experiments. However, the advantages of using naturally complex and metabolically diverse populations overshadow this latent error. The applied community-level toxicity testing approach takes into account both the fate and toxicity of the test chemical in an integrating fashion and should be seen as a first step when aiming to use microbial communities for the specific purpose of toxicity testing. The lack of higher trophic levels is expected to facilitate data interpretation as compared to mesocosm or field studies, and the proposed microcosm ap-

proach should thus make it possible to identify sensitive subpopulations of the tested microbial community for further toxicity testing under more field-like conditions. Based on our study, application of a polyphasic approach, targeting both the structure and function of the microbial communities, and including a suite of overlapping methodologies should be applied to facilitate data interpretation. Specifically we recommend that community-level toxicity testing should be performed using sensitive effect indicators such as PCR-DGGE or related DNA fingerprinting techniques and assays for assessment of the actual growth rate (e.g., incorporation of leucine). These combined approaches will allow sensitive detection of changes in both structure and function of the microbial community.

Acknowledgements Regitze E. Jensen, Ann-Siri Borg Hentze, and Anne Grethe Holm-Jensen are acknowledged for technical assistance. The study was supported by the Danish National Sciences Research Council and by the Danish Environmental Research Program (Center for Sustainable Land Use and Management of Contaminants, Carbon, and Nitrogen).

References [1] van Leeuwen, C.J. (1996) General introduction. In: Risk Assessment of Chemicals: An Introduction (van Leeuwen, C.J. and Hermens, J.L.M., Eds.), pp. 1–18. Kluwer Academic, London, UK. [2] Kimerle, R.A. (1989) Aquatic and terrestrial ecotoxicology of linear alkylbenzene sulfonate. Tenside Surfact. Det. 26, 169–176. [3] Amann, R., Ludwig, W. and Schleifer, K.-H. (1995) Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol. Rev. 59, 143–169. [4] Guckert, J.B. (1996) Toxicity assessment by community analysis. J. Microbiol. Meth. 25, 101–112. [5] Blanck, H. (2002) A critical review of procedures and approaches used for assessing pollution-induced community tolerance (PICT) in biotic communities. Hum. Ecol. Risk Assess. 8, 1003–1034. [6] Ford, T.E. (2000) Response of marine microbial communities to anthropogenic stress. J. Aquat. Ecosyst. Stress Recov. 7, 75–89. [7] Atlas, R.M., Horowitz, A., Krichevsky, M. and Bej, A.K. (1991) Response of microbial populations to environmental disturbance. Microb. Ecol. 22, 249–256. [8] Cavalli, L., Clerici, R., Radici, P. and Valtorta, L. (1999) Update on LAB/LAS. Tenside Surfact. Det. 36, 254–258. [9] Painter, H.A. (1992) Anionic surfactants. In: Detergents (de Oude, N.T., Ed.). Springer, Berlin. [10] Feijtel, T., Vits, H., Murray-Smith, R., van Wijk, R., Koch, V., Schr€ oder, R., Birch, R. and Berge, W.T. (1996) Fate of LAS in activated sludge wastewater treatment plants: a model verification study. Chemosphere 32, 1413–1426. [11] Le on, V.M., Saez, M., Gonzalez-Mazo, E. and Gomez-Parra, A. (2002) Occurrence and distribution of linear alkylbenzene sulfonates and sulfophenylcarboxylic acids in several Iberian littoral ecosystems. Sci. Tot. Environ. 288, 215–226.

K.K. Brandt et al. / FEMS Microbiology Ecology 49 (2004) 229–241 [12] Marcomini, A., Di Corcia, A., Samperi, R. and Capri, S. (1993) Reversed-phase high-performance liquid chromatographic determination of linear alkylbenzene sulphonates, nonylphenol polyethoxylates and their carboxylic biotransformation products. J. Chromatogr. A 644, 59–71. [13] Iizumi, T., Mizumoto, M. and Nakamura, K. (1998) A bioluminescence assay using Nitrosomonas europaea for rapid and sensitive detection of nitrification inhibitors. Appl. Environ. Microbiol. 64, 3656–3662. [14] Brandt, K.K., Hesselsøe, M., Roslev, P., Henriksen, K. and Sørensen, J. (2001) Toxic effects of linear alkylbenzene sulfonate on metabolic activity, growth rate, and microcolony formation of Nitrosomonas and Nitrosospira strains. Appl. Environ. Microbiol. 67, 2489–2498. [15] Brandt, K.K., Pedersen, A. and Sørensen, J. (2002) Solid phasecontact assay that uses a lux-marked Nitrosomonas europaea reporter strain to estimate toxicity of bioavailable linear alkylbenzene sulfonate in soil. Appl. Environ. Microbiol. 68, 3502– 3508. [16] Jørgensen, N.O.G. (1992) Incorporation of [3 H]-leucine and [3 H]valine into proteins of freshwater bacteria: field applications. Appl. Environ. Microbiol. 58, 3647–3653. [17] Tulonen, T. (1993) Bacterial production in a mesohumic lake estimated from [14 C]-leucine incorporation rate. Microb. Ecol. 26, 201–217. [18] Kirchman, D.L. (1993) Leucine incorporation as a measure of biomass production by heterotrophic bacteria. In: Handbook of Methods in Aquatic Microbial Ecology (Kemp, P.F., Sherr, B.F., Sherr, E.B. and Cole, J.J., Eds.), pp. 509–512. Lewes, New York. [19] Smith, D.C. and Azam, F. (1992) A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3 H-leucine. Mar. Microb. Food Webs 6, 107–114. [20] Simon, M. and Azam, F. (1989) Protein content and protein synthesis rates of planktonic marine bacteria. Mar. Ecol. Prog. Ser. 51, 201–213. [21] Tezuka, Y. (1990) Bacterial regeneration of ammonium and phosphate as affected by the carbon:nitrogen:phosphorus ratio of organic substrates. Microb. Ecol. 19, 227–238. [22] Hobbie, J.E., Daley, R.J. and Jasper, S. (1977) Use of nuclepore filters for counting bacteria by fluorescence microscopy. Appl. Environ. Microbiol. 33, 1225–1228. [23] Christoffersen, K., Lyck, S. and Winding, A. (2002) Microbial activity and bacterial community structure during degradation of microcystins. Aquat. Microb. Ecol. 27, 125–136. [24] Insam, H. (1997) A new set of substrates proposed for community characterization in environmental samples. In: Microbial Communities: Functional versus Structural Approaches (Insam, H. and Rangger, A., Eds.), pp. 259–260. Springer, Berlin. [25] Garland, J.L. (1996) Analytical approaches to the characterization of samples of microbial communities using patterns of potential carbon source utilization. Soil Biol. Biochem. 28, 213–221. [26] Garland, J.L. (1997) Analysis and interpretation of communitylevel physiological profiles in microbial ecology. FEMS Microbiol. Ecol. 24, 289–300.

241

[27] Garland, J.L., Mills, A.L. and Young, J.S. (2001) Relative effectiveness of kinetic analysis vs single point readings for classifying environmental samples based on community-level physiological profiles (CLPP). Soil Biol. Biochem. 33, 1059–1066. [28] Rutgers, M., van’t Verlaat, I.M., Wind, B., Posthuma, L. and Breure, A.M. (1998) Rapid method for assessing pollutioninduced community tolerance in contaminated soil. Environ. Toxicol. Chem. 17, 2210–2213. [29] Siciliano, S.D., Gong, P., Sunahara, G.I. and Greer, C.W. (2000) Assessment of 2,4,6-trinitrotoluene toxicity in field soils by pollution-induced community tolerance, denaturing gradient gel electrophoresis, and seed germination assay. Environ. Toxicol. Chem. 19, 2154–2160. [30] Gong, P., Siciliano, S.D., Srivastava, S., Greer, C.W. and Sunahara, G.I. (2002) Assessment of pollution-induced microbial community tolerance to heavy metals in soil using ammonia-oxidizing bacteria and biolog assay. Hum. Ecol. Risk Assess. 8, 1067–1081. [31] Riemann, L. and Winding, A. (2001) Community dynamics of free-living and particle-associated bacterial assemblages during a freshwater phytoplankton bloom. Microb. Ecol. 42, 274–285. [32] Jørgensen, E. and Christoffersen, K. (2000) Short-term effects of linear alkylbenzene sulfonates on freshwater plankton studied under field conditions. Environ. Toxicol. Chem. 19, 904–911. [33] Martinez, J., Vives-Rego, J. and Sanchez-Leal, J. (1989) The effect of chemical structure and molecular weight of commercial alkylbenzenes on the toxic response of Daphnia and naturally occurring bacteria in fresh and seawater. Water Res. 23, 569–572. [34] Belanger, S.E., Bowling, J.W., Lee, D.M., LeBlanc, E.M., Kerr, K.M., McAvoy, D.C., Christman, S.C. and Davidson, D.H. (2002) Integration of aquatic fate and ecological responses to linear alkyl benzene sulfonate (LAS) in model stream ecosystems. Ecotox. Environ. Safe. 52, 150–171. [35] Lee, C., Russell, N.J. and White, G.F. (1995) Modelling the kinetics of biodegradation of anionic surfactants by biofilm bacteria from polluted riverine sites: a comparison of five classes of surfactant at three sites. Water Res. 29, 2491–2497. [36] Breen, A., Jimenez, L., Saylor, G.S. and Federle, T.W. (1992) Plasmid incidence and linear alkylbenzene sulfonate biodegradation in wastewater and pristine pond ecosystems. J. Indust. Microbiol. 9, 37–44. [37] Lee, C., Russell, N.J. and White, G.F. (1995) Rapid screening for bacterial phenotypes capable of biodegrading anionic surfactants: development and validation of a microtitre plate method. Microbiol. 141, 2801–2810. [38] Hrsak, D. and Begonja, A. (1998) Growth characteristics and metabolic activities of the methanotroph-heterotrophic groundwater community. J. Appl. Microbiol. 85, 448–456. [39] Jimenez, L., Breen, A., Thomas, N., Federle, T.W. and Sayler, G.S. (1991) Mineralization of linear alkylbenzene sulfonate by a four member aerobic bacterial consortium. Appl. Environ. Microbiol. 57, 1566–1569. [40] Sigoillot, J.-C. and Nguyen, M.-H. (1992) Complete oxidation of linear alkylbenzene sulfonate by bacterial communities selected from coastal seawater. Appl. Environ. Microbiol. 58, 1308–1312.