Effect of plants and filter materials on bacteria removal in pilot-scale constructed wetlands

ARTICLE IN PRESS

Water Research 39 (2005) 1361–1373 www.elsevier.com/locate/watres

Effect of plants and filter materials on bacteria removal in pilot-scale constructed wetlands Gabriela Vaccaa, Helmut Wandb, Marcell Nikolausza, Peter Kuschka, Matthias Ka¨stnera, a

Department of Bioremediation, UFZ Centre for Environmental Research Leipzig-Halle, Permoserstr. 15, D-04318 Leipzig, Germany b Saxonian Institute for Applied Biotechnology, Permoserstr. 15, D-04318 Leipzig, Germany Received 16 September 2004; received in revised form 3 January 2005; accepted 7 January 2005 Available online 17 February 2005

Abstract Due to the lack of testing units or appropriate experimental approaches, only little is known about the removal of bacteria in constructed wetlands. However, improved performance in terms of water sanitation requires a detailed understanding of the ongoing processes. Therefore, we analyzed the microbial diversity and the survival of Enterobacteriaceae in six pilot-scale constructed wetland systems treating domestic wastewater: two vertical sand filters, two vertical expanded clay filters and two horizontal sand filters (each planted and unplanted). Samples were taken from the in- and outflow, from the rhizosphere, and from the bulk soil at various depths. Colony-forming units of heterotrophic bacteria and coliforms were analyzed and the removal of bacteria between the in- and outflow was determined to within 1.5–2.5 orders of magnitude. To access the taxon-specific biodiversity of potential pathogens in the filters and to reduce the complexity of the analysis, specific primers for Enterobacteriaceae were developed. While performing PCR–SSCP analyses, a pronounced decrease in diversity from the inflow to the outflow of treated wastewater was observed. No differences were observed between the bulk soil of planted and unplanted vertical filters. Some bands appeared in the rhizosphere that were not present in the bulk soil, indicating the development of specific communities stimulated by the plants. The fingerprinting of the rhizosphere of plants grown on sand or expanded clay exhibited many differences, which show that different microbial communities exist depending on the soil type of the filters. The use of the taxon-specific primers enabled us to evaluate the fate of the Enterobacteriaceae entering the wetlands and to localize harboring in the rhizosphere. The most abundant bands of the profiles were sequenced: Pantoea agglomerans was found in nearly all samples from the soil but not in the effluent, whereas Citrobacter sp. could not be removed by the horizontal unplanted sand and vertical planted expanded clay filters. These results show that the community in wetland system is strongly influenced by the filtration process, the filter material and the plants. r 2005 Elsevier Ltd. All rights reserved. Keywords: Constructed wetlands; SSCP; Community analysis; Enterobacteriaceae; Bacteria removal; Subsurface flow CWs

1. Introduction

Corresponding author. Tel.: +49 341 235 2746;

fax: +49 341 235 2492. E-mail address: [email protected] (M. Ka¨stner).

Health risks rise sharply with the ingestion of unsafe water: diseases related to water sanitation are estimated to account for 4.0% of all deaths and 5.7% of the total disease burden occurring worldwide (Pru¨ss et al., 2002).

0043-1354/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2005.01.005

ARTICLE IN PRESS 1362

G. Vacca et al. / Water Research 39 (2005) 1361–1373

Therefore, research into sewage treatment is needed in order to reduce the risks associated with improper sanitation, particularly in terms of wastewater reuse for crop irrigation. One way of treating wastewater is to apply constructed wetlands, an easy to handle technique that is considered to be relatively inexpensive in developing countries (Mashauri et al., 2000). Many studies have been focused on removing the organic load and other contaminants from sewage (Yang et al., 2001; Hench et al., 2003), stormwater (Green et al., 1997), industrial wastewater (Vrhovsek et al., 1996), agricultural runoff (Higgins et al., 1993), acid mine drainage (Karanthanasis and Thompson, 1995) and landfill leachate (Bernard and Lauve, 1995) using constructed wetlands. However, limited attention has been paid to examining the efficacy of constructed wetlands for the removal of enteric microorganisms from domestic sewage. Studies that have been documented (Gersberg et al., 1989a, b; Rivera et al., 1995; Williams et al., 1995; Green et al., 1997) are usually limited to total and/or faecal coliform removal simply measured in terms of colony-forming units or most probable number techniques in samples from the in- and outflow of pilot- or full-scale wetlands. These techniques are highly sensitive, well established and standardized, but provide information only on the presence or absence of certain bacteria and not on the community structure. Constructed wetlands receiving primary municipal wastewater were found to reduce the total numbers of coliforms in the effluent of planted filter beds from a level of 6.7
105 ml1 by 99% (Gersberg et al., 1989a). Other authors found a removal of Escherichia coli of 31–91% in the rhizosphere but only 0–35% in the unvegetated controls (Rivera et al., 1995). Subsurfaceflow constructed wetlands exhibited removals of E. coli of 1.5–2.1 orders of magnitude (Green et al., 1997). Removals of 2.4–5.3 orders of magnitude for cultivable Salmonella cells were found (Pundsack et al., 2001). A linear correlation was found between the reduction of coliforms and the flow distance along a horizontal flow gravel bed system fed with primary and secondary sewage effluent (Williams et al., 1995). At higher hydraulic loading rates of 520–1300 mm/d, a removal of only 0.7–1.5 orders of magnitude of total coliforms and faecal streptococci was observed in a vertical-flow wetland system (Arias et al., 2003) and a strongly negative correlation of bacterial removal with the flow rate through the wetland systems was observed (Perkins and Hunter, 2000). In terms of bacteria removal, constructed wetlands are generally considered to feature a combination of chemical and physical factors, including mechanical filtration and sedimentation (Pundsack et al., 2001). Chemical factors include oxidation, exposure to biocides excreted by a number of plants, and sorption to organic

matter. The biological removal mechanisms may include antimicrobial activity of root exudates (Kickuth and Kaitzis, 1975; Axelrood et al., 1996), predation by nematodes and protists (Decamp and Warren, 1998; Decamp et al., 1999), activity of lytic bacteria or viruses (Axelrood et al., 1996), retention in biofilms (Brix, 1997), and natural die-off (Gersberg et al., 1989a, b). Even though possible mechanisms of bacterial removal have been discussed in many papers (Burger and Weise, 1984; Armstrong et al., 1990; Morales et al., 1996; Decamp and Warren, 2000), no systematic analyses on the removal processes and the fate of potential pathogenic bacteria in constructed wetlands are yet known. In particular, the structure of microbial populations in response to vertical or horizontal subsurfaceflow systems, soil type, plant presence, and BOD load, etc. is not understood. Within the last decade, molecular fingerprinting techniques have been applied in order to gain an easier access to the structure of bacterial communities of various ecosystems (Amann et al., 1995; Muyzer and Smalla, 1998; Bull et al., 2000). However, only one report was found in which the microbial communities of two constructed wetlands for the treatment of agricultural waste were analyzed in terms of total bacterial community and ammonia-oxidizing bacteria composition (Ibekwe et al., 2003). Due to the complexity of the microbial communities detected with general primers, more restricted approaches with taxon-specific primers are required and were recently followed for the detection of potential pathogenic bacteria such as Campylobacter and Yersinia species (Alexandrino et al., 2004). However, these methods usually rely on the amplification of virulence-associated or antibiotic-resistance genes (Dutka-Malen et al., 1995), or else taxon-specific detection is restricted to the genus level (Yersinia) (Lantz et al., 1998). In order not only to evaluate the removal efficiency of wastewater treatment but also to assess the changes in the community structure, molecular fingerprinting approaches are needed. Although taxonspecific probes for enteric bacteria have already been developed for FISH detection (Friedrich et al., 2003), the direct application of these probes to PCR-based molecular fingerprinting techniques, such as DGGE or single-strand conformation polymorphism (SSCP), are generally not straightforward, therefore the development of new primers was necessary. The goal of the present work was to analyze the removal of bacteria and the changes of the microbial community structure in constructed wetlands. Therefore, the effects of the process scheme (vertical or horizontal subsurface-flow wetland systems), the presence of plants and the soil filter material were determined in six pilot-scale plants operating with real domestic wastewater in parallel. In order to reduce complexity of the analyses, molecular fingerprinting

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

1363

2. Materials and methods

soil or 100 ml of water was taken at each sampling site. Samples were divided in aliquots for each parameter. Soil samples were sieved (2 mm mesh size) in order to remove organic debris and larger inorganic fragments. Each sample was mixed and then divided into aliquots for the various analyses. Samples were collected in sterile flasks. Roots were separated carefully from bulk soil and stored in separate sterile containers. All samples were transported and stored immediately at 4 1C for microbiological parameters and 20 1C for DNA extraction.

2.1. Constructed wetlands and sampling conditions

2.2. Extraction of bacteria cells and cultivation conditions

The wetlands were located at the wastewater treatment facility in Langenreichenbach, Germany. Technical details about the construction and operation of the wetland system were described previously (BaederBederski et al., 2004). Briefly, the open-air constructed wetlands consisted of 14 filter beds (plots) with a surface area of 6.72 m2 each, which received raw wastewater with a daily hydraulical load of 60 l m2 d1 (60 mm d1). This unique pilot plant has the advantage that every plot definitely received the same wastewater inflow and showed in 2002 an averaged COD removal of about 90%. Therefore, we decided to perform a crosscutting sampling approach at a certain date for comparison of the different treatment processes. Triplicate samples were taken in a sampling campaign in August 2002 from the inflow, outflow, different depths and compartments (rhizosphere and bulk soil) of six constructed wetland systems: two vertical-flow sand filters, two vertical-flow expanded clay filters, and two horizontal-flow sand filters (planted with Phragmites australis and unplanted variants of each type) (Table 1). The soil samples (bulk and rhizosphere soil) from the vertical-flow filters were collected at depths of 15, 35 and 45 cm and from the horizontal-flow filters at depths of 10, 20 and 30 cm at a distance of 1.5, 2.5, 3.5 and 4.5 m away from the inflow using a Purkhauer auger (diameter 30 mm) or an Edelman auger (Fig. 1). Hundred grams of

Soil samples (1 g each) were resuspended in 10 ml sodium pyrophosphate solution (0.2%, wt/vol) and sonicated for 7 min. After sedimentation for 30 min at room temperature, the samples were immediately diluted in sterile saline solution (0.85% [wt/vol] NaCl) and inoculated in triplicate onto appropriate media. The roots were vigorously washed in saline solution several times. Rhizosphere fractions were obtained by pelleting the washing solutions by centrifugation (12,000g, 10 min). The pellets were suspended in 10 ml portions of pyrophosphate solution (0.2%, wt/vol) in Falcon tubes and treated in the same way as the soil samples. The resulting dilutions, the outflows of the various plots and the inflow of domestic wastewater were analyzed using a modified MPN method based on plate-counting on R2A agar (Difco, US) for total CFU and on Endo agar (Merck, Germany) for coliform bacteria. Serial dilutions in a variable range of 100–106 were each dotted with a volume of 10 ml in triplicates onto a single plate of the corresponding agar medium. In the case of low cell concentrations, 100 ml of the undiluted sample were plated. Colonies within the dots of the plates were counted and the CFUs were calculated for the parent samples and statistically analyzed using the known procedures for MPN methods (APHA, 1998). Standard deviations using this method were o10%.

methods were used and a new specific primer set for the taxon Enterobacteriaceae was developed in order to trace the fate of this group of potential pathogenic bacteria. Samples were taken from the inflow, rhizosphere and bulk soil at various distances from the inflow in order to gain insight into the ‘‘black box’’ of the wetland soil body.

Table 1 Sampling sites and characteristics of various plots of the pilot-scale constructed wetlands in Langenreichenbach, Germany Plot

System

Plant

Filter material

Depth (cm)

HPS HUS VPS VUS VPC VUC

Horizontal Horizontal Vertical Vertical Vertical Vertical

Phragmites australis Unplanted Phragmites australis Unplanted Phragmites australis Unplanted

Sanda Sand Sand Sand Expanded clay/sand Expanded clay/sand

10, 10, 15, 15, 15, 15,

H ¼ horizontal; V ¼ vertical; U ¼ unplanted; P ¼ planted; S ¼ sand; C ¼ expanded clay. a Grain size 0–2 mm. b Samples taken at four different distances from inflow: 1.5, 2.5, 3.5 and 4.5 m.

20, 20, 35, 35, 35, 35,

30b 30b 45 45 45 45

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

1364

2.4 m 1.0 m

Inflow

Sampling sites

15 cm 35 cm 45 cm

2.8 m

Outflow

(A)

Inflow

10 cm

1.2 m 0.6 m

20 cm

5.6 m

30 cm

Outflow

Saturated zone 1.5 m 2.5 m 3.5 m

(B)

4.5 m

Fig. 1. Sampling sites in various planted or unplanted plots of the constructed wetland pilot plant: (A) vertical filters and (B) horizontal filters. Flow directions are indicated by arrows and the size of the filters is indicated by the x, y and z-axis on the left side. Black cylinders represent the augers used for taking samples of the rhizosphere.

2.3. DNA extraction Community DNA was obtained from the bulk soil and from the rhizosphere fraction. DNA extracts were obtained using the FastDNA SPIN soil kit with the protocol recommended by the manufacturer (Qbiogene, Carlsbad, CA, USA). DNA extractions were performed in duplicate. 2.4. Primer design and PCR amplification of partial SSU rRNA genes A new taxon-specific primer (Ent18-Ph primer, 50 GCA ACA AAG GAT AAG GGT 30 ) was developed, which binds to a highly specific region of the SSU rRNA gene of Enterobacteriaceae, and applied as a reverse primer in combination with Com1 in order to assess the community structure, distribution, and the fate of members of this group of potential pathogens. The primer was designed by the Probe Design Tool of ARB (Ludwig et al., 2004) and the specificity of this sequence was examined further by comparing it to the National Centre for Biotechnology Information (NCBI) sequence databases by using the BLAST function (Altschul et al., 1998). The sequence was also analyzed using the online Probe Match program of the Ribosomal Database Project II (Maidack et al., 1997). The optimum

annealing temperature for specific amplification was 56 1C, which was tested with E. coli DSM 498, Campylobacter jejuni subsp. jejuni LMC 5, Yersinia enterocolitica DSM 11502 and Helicobacter pylori CCUG 17,874 strains. Bacillus subtilis DSM10 and Enterococcus faecalis OGIX strains were used as negative controls. The primers were used to amplify 16S rRNA genes from nucleotide 519 to nucleotide 1127 (E. coli numbering). 16S rRNA gene sequences were also amplified using primers Com1 and Com2-Ph (Schwieger and Tebbe, 1998). For the SSCP analysis, forward primers were phosphorilated at the 50 end to obtain single stranded DNA by lambda nuclease digestion. Each PCR was performed using a total volume of 100 ml in a micro test tube. The reaction mixtures each contained 1X PCR buffer, 200 mM of each dNTP, 0.5 mM of each primer, 4 ml of DNA extract from the sample to be tested and 2.5 U of HotStar Taq polymerase (Qiagen, Hilden, Germany). Amplification reactions were performed in a Mastercycler gradient (Eppendorf, Hamburg, Germany). An initial denaturation step at 95 1C for 15 min was followed by 35 cycles each consisting of 60 s at 94 1C, 60 s at 50 1C and 70 s at 72 1C followed by a final extension at 721 for 5 min. PCR products were separated on a 1.5% agarose gel stained with ethidium bromide and visualized with UV excitation.

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

2.5. Genetic profiles by SSCP PCR products were purified with Qiaquick columns using a protocol recommended by the manufacturer (Qiagen). Samples were eluted with 30 ml of Tris-HCl (pH 8.0) buffer supplied by the manufacturer. For the digestion of the phosphorylated strand, 40 U of lambda exonuclease (New England Biolabs, Schwalbach, Germany) was mixed with 300 ng of the eluted PCR products in a 30 ml total volume mixture containing 1X (final concentration) lambda exonuclease reaction buffer (New England Biolabs). The reaction mixtures were incubated at 37 1C for 45 min, purified with MiniElute columns (Qiagen), and eluted with 10 ml of Tris-HCl buffer. For electrophoresis, 9 ml of denaturating loading buffer (95% [vol/vol] formamide, 10 mM NaOH, 0.25% [wt/vol] bromophenol blue, 0.25% [wt/vol] 0.25% xylene cyanol) was added to 9 ml of the single-stranded DNA solutions. Samples were incubated at 95 1C for 2 min and immediately cooled on ice. After 3 min, samples were loaded onto the gel in equal amounts. The samples were separated by electrophoresis on a 0.625
MDE gel (FMC Bioproducts, Rockland, ME, USA) with 1
TBE buffer. The gels were cast using 0.5 mm spacers and a thermostatic plate as recommended by the manufacturer. The gels (length 20 cm) were run at 400 V for 16 h at 20 1C with a Protean II xi cell (Biorad, Munich, Germany). Gels were fixed to the front glass plate with Bind Silane (Amersham Pharmacia Biotech, Freiburg, Germany) and silver-stained according to Bassam et al. (1991). The GelCompare suite software package (version 2.5; A, Applied Maths, St-Martens-Latem, Belgium) was used to analyze the similarity of SSCP patterns on each gel. Pearson correlation coefficients and the unweighted pair group method using arithmetic averages (UPGMA) clustering algorithm was used to construct matrices and dendrograms. 2.6. Identification of SSCP bands Selected prominent bands detected in MDE polyacrylamide gels after silver-staining were excised with razor blades, and single-stranded DNA was eluted from the gel with 40 ml distilled water overnight at 4 1C. The single-strand molecules were re-amplified by PCR using the same primers and conditions as for the respective analysis. The resulting PCR products were analyzed by SSCP for purity and identity by comparing them with the original fragments in the community profiles. Purified PCR products (2 ml) were ligated into a pGEM-TEasyTM vector and transformed into JM109 high efficiency competent cells (Promega, Madison, WI) as described by the manufacturer. Clone libraries (50 clones of each) were screened as described earlier (Nikolausz et al., 2004) and only two members of the

1365

dominant restriction group were sequenced. Sequencing reactions of both strands with M13(20) and M13rev primers were carried out using the ABI PRISM BigDye Terminator Cycle Sequencing Kit V. 3.0 (Applied Biosystems) according to the manufacturer’s protocol. Sequencing reaction mixtures were separated with an ABI PRISM 3730 DNA Analyzer (Applied Biosystems, Foster City, CA). Analysis of sequences and homology searches were performed using the BLAST algorithm (Altschul et al., 1998) with the BLAST server of the National Centre for Biotechnology Information. Nucleotide sequence data reported are available in the EMBL database under the accession numbers from AJ832130 to AJ832137.

3. Results and discussion 3.1. Removal of Enterobacteriaceae in constructed wetlands In order to analyze the quantitative removal of bacteria from wastewater, the quantification of bacteria using CFU methods is appropriate; these methods are well established (APHA, 1998). Therefore, the inflow of domestic wastewater and the outflows of the various plots of the pilot-scale constructed wetlands were analyzed with total plate count methods on R2A agar and for coliform bacteria on Endo agar. All filters received the same inflow of wastewater and were operated with the same climate and hydraulic load. The C and N removal efficiency of the pilot plants was analyzed over a period of 3 years; in general, sand filters performed better than expanded clay plots, vertical filter plots better than horizontal filter plots and unplanted filters in some cases as well or even better than planted filters (Baeder-Bederski et al., 2004). Table 2 Total and coliform bacteria CFU in the inflow and outflow samples Inflow/plots

Coliform [Log (cfu (100 ml1)]

Inflow PFb HPS HUS VPS VUS VPC VUC

8.2 8.2 6.1 5.7 5.9 5.7 6.8 6.5

a

(0.1)a (0.1) (0.1) (0.1) (0.1) (0.2) (0.1) (0.1)

Heterotrophic [Log (cfu (100 ml1)] 8.3 8.3 6.5 6.4 6.5 6.3 7.1 7.4

(0.1) (0.1) (o0.1) (o0.1) (o0.1) (0.1) (0.2) (0.2)

Standard deviation. Primary filtration of the inflow, particle elimination before inflow of the various plots. b

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

sample. Fig. 2 shows SSCP profiles of PCR-amplified partial 16S rRNA gene sequences of DNA extracted from bulk soil of the horizontal planted sand filter depending on the distance from the inflow and the depth. Some differences were detected in these profiles (Fig. 2, bands b–d), but the diversity was too high for an effective assessment of the removal processes. Therefore, a new taxon-specific primer set for the detection of Enterobacteriaceae was developed in order to improve the resolution of the profiles and to monitor the relevant microbes for hygienic aspects. From a hygienical point of view, this taxon is most relevant, since it represents the predominant bacteria in wastewater and most of the pathogenic bacteria or facultative pathogens belong to this group. DNA sample amplifications with the developed Com1/Ent18 primers were highly reproducible and generated products which included a highly specific region of the SSU rRNA gene of Enterobacteriaceae. The specificity of the primer was

The number of coliform bacteria was generally reduced by two orders of magnitude in all plots (Table 2, see Table 1 for description of the various plots). The observed removal rates (499%) are in accordance with known results from other full-scale wetland systems (Gersberg et al., 1989a, b; Rivera et al., 1995; Green et al., 1997; Arias et al., 2003). However, the differences between the values in the outflows of the various plots indicate that the removal efficiency depended on the presence of plants, the filter material and the conditions of operation. CFU-based methods provide quantitative data on the removal of bacteria but no insight into the community structure and the changes during the treatment processes. Hence, a molecular–biological approach was applied in order to gain insight into the community structure of the various plots. Using the universal 16S rDNA primers (Com1/Com2), the corresponding 407 bp fragment of the SSU rRNA was obtained from each

Outlfow

1366

a b c

a b c

a b c

d f

d e

d f

g

b c d e g

a b c

b c

d f

d e g

h

h

1.5

2.5

h

3.5

30

10

30

20

10

30

20

10

30

20

10

Depth [ cm]

a b c

a b c

a b c

d f g

d f g

d e

h

h

h

f

g

4.5

Distance from inflow [ m ] Fig. 2. SSCP profiles of PCR-amplified partial 16S rRNA gene sequences of DNA extracted from bulk soil of the horizontal planted sand filter depending on the distance from inflow and the depth. Products generated by universal primers (com1/com2) include variable regions 4 and 5. Characters indicate bands which (i) are in most of the samples (band a); (ii) appear (band b, c, e and h) or (iii) disappear (band d) with increasing depth.

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

(Normander and Prosser, 2000; Berg et al., 2002), soil, water, as well as with humans and other animals (Gavini et al., 1989). Knowledge of the habitat of Citrobacter is based on its occurrence as an intestinal commensal. It is reported to occur in water, sewage and soil (Sedlak, 1973) as well as in the rhizosphere (Zablotowicz et al., 1995). S. maltophilia is a Gram-negative bacillus that is common in the rhizosphere environment (Berg et al., 1996; Lottmann et al., 1999). However, it is from a taxon closely related to Enterobacteriaceae. The outflow of the vertical sand filter provided the lowest diversity while the horizontal unplanted sand filter had the lowest quantity of enteric bacteria and the highest diversity expressed in terms of the number of the SSCP bands (Fig. 3. and Table 2). Thus, not only the monitoring of removal efficiency in terms of CFU should be taken into account but also the quality of bacteria reduction, because some bands were removed from the profiles while others are enriched depending on the characteristics of the filter: the band representing Pantoea was completely removed in both planted sand systems; bands of Citrobacter remained in the planted expanded clay system and disappeared from other filters; Stenotrophomonas appears to be present in planted sand systems with very low intensity and is enriched in unplanted sand systems but not in filters with expanded clay; and Rheinheimera-like bacteria were enriched in horizontal planted sand filters and to a lesser extent in vertical planted expanded clay filters but disappeared in other vertical systems.

VUC

VPC

VUS

VPS

HUS

HPS

PF

tested on DNAs isolated from strains belonging to Enterobacteriaceae and to other taxa. Although under sub-optimal annealing temperatures this primer pair also amplified non-Enterobacteriaceae sequences from pure cultures, under optimal conditions we obtained only specific PCR products from DNA isolated from pure cultures. However, from environmental samples we cannot exclude amplification of other closely related taxa. Fig. 3 shows the SSCP profiles of the amplified 16S rRNA gene sequences of the bacterial consortia extracted from the inflow wastewater and from the outflows of the various filter plots. Although the efficiency of the bacteria removal from the outflow of planted and unplanted plots was similar in terms of CFU, the composition of the bacterial community as resolved by SSCP with the new primers differed greatly. This demonstrates the effectivity of the taxon-specific approach for gaining insight into the changes of the community during the various wetland treatment processes. The most prominent bands appearing in the inflow or outflow of most of the filter plots revealed the persistence of certain enteric organisms that were identified by sequencing as Pantoea agglomerans, Citrobacter freundii, and genospecies closely related to Stenotrophomonas maltophila and distantly related to Rheinheimera sp. (o97%; Fig. 3). Members of the genus Pantoea are ubiquitous in nature, occurring in many different habitats, including in association with plants

Inflow

1367

Pantoea agglomerans (100% similarity) Stenotrophomonas maltophilia (99% similarity)

Citrobacter freundii (100% similarity) Rheinheimera sp. (97% similarity)

Fig. 3. SSCP profiles of PCR-amplified partial 16S rRNA gene sequences using Enterobacteriacea-specific priming (primers Com1/ Ent18). DNA was extracted from wastewater inflow, after primary filtration (PF) and from the outflows of the various plots. All labelled bands were excised from the gel, re-amplified and subjected to sequence analysis. Closest relatives are indicated together with the obtained similarity values.

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

1368

3.2. Effect of the plants

Log (cfu g-1)

The overall removal efficiency in terms of CFU of heterotrophic bacteria in the outflows of the filter plots was about two orders of magnitude, irrespective of the planted/non-planted or vertical/horizontal nature of the plots with sand, whereas the plots with expanded clay showed a significantly lower level of removal (Table 2). However, no correlation was found between the removal efficiency in terms of enteric CFU and the presence of the plants. The removal efficiencies of the filter plots decreased as follows: horizontal unplanted sand4vertical unplanted sand4vertical planted sand4horizontal planted sand4vertical unplanted expanded clay4vertical planted expanded clay. Irrespectively of the substrate, the value of heterotrophic CFU and even of coliforms found in the 9 8 7 6 5 4

9 8 7 6 5 4

3 2

3 2

Log (cfu g-1)

1 0 -1

1 0 -1

A,10

9 8 7 6 5 4

9 8 7 6 5 4

3 2

3 2

1 0 -1

Log (cfu g-1)

rhizosphere of the filters was at least two orders of magnitude higher than in the bulk soil (Figs. 4 and 5). Previous studies showed that the bacterial populations are generally enhanced in the rhizosphere (Brix, 1997; Chiarini et al., 1998; Soto et al., 1999). In addition, our results show that enteric bacteria appear to form a part of the rhizosphere communities in the constructed wetland, and that the exudates of Phragmites australis or other competing bacteria in its rhizosphere (Axelrood et al., 1996; Pierson and Pierson, 1996) do not cause a significant elimination of this group. The effect of plants was demonstrated by statistical cluster analysis on the basis of the SSCP fingerprintings from vertical sand filters (Fig. 6). The separation of the rhizosphere and bulk soil samples clearly indicates the rhizosphere effect. This result shows that the plants ‘stimulate’ the formation of certain communities in the

1 0 -1

A,20

9 8 7 6 5 4

9 8 7 6 5 4

3 2

3 2

1 0 -1

1 0 -1

A,30 1.5

2.5

3.5

B,10

B,20

B,30

4.5 1.5 Distance from inflow [m]

2.5

3.5

4.5

Fig. 4. Heterotrophic (A) and coliform (B) bacteria in horizontal filters. Depths: 10; 20; 30 cm —m—: HPS bulk soil; —’—: HUS bulk soil; ?J?: HPS rhizosphere.

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373 9

8

8

7

7

6

6

5

5

(A) 4

4

9

9

8

8

7

7

6

6

5

5

Log (cfu g-1)

Log (cfu g-1)

9

4 15

35

(B)

4 45 15 Depth [cm]

1369

35

45

Fig. 5. Heterotrophic (A) and coliform (B) bacteria in vertical filters in relation to depth: —m— VPS bulk soil; —.— VUS bulk soil; —,— VPS rhizosphere; —J—: VPC rhizosphere; —K—: VPC bulk soil; ?K?: VUC bulk soil.

75

80

85

90

95 100 VUS -bulk soil-15 cm VUS -bulk soil-45 cm VUC -bulk soil-15 cm VPC -bulk soil-35 cm VPC -bulk soil-15 cm VPC -bulk soil-45 cm VUC -bulk soil-45 cm VUC -bulk soil-35 cm VPS -bulk soil-45 cm VPS -bulk soil-35 cm VPS -bulk soil-15 cm VPC –rhizosphere-45 cm VPC –rhizosphere-35 cm VPC –rhizosphere-15 cm VPS –rhizosphere-15 cm

Fig. 6. Cluster analysis of SSCP profiles with the Enterobacteriaceae-specific primers (Com1/Ent18) at various depths of vertical filters.

wastewater, which is in good agreement with previous results (Smalla et al., 2001). However, no amplification with the primers Com1/ Ent18 was observed in the bulk soil of the horizontal filter plots; the lack of PCR products due to the presence of inhibiting substances could be definitely excluded. The lack of PCR products compared to the coliform

CFU values of the same samples is an indirect indication of the sensitivity of this PCR approach. Although this sensitivity is not comparable to the most advanced platecounting methods in microbiology, the information obtained about the changes in the community composition by our PCR–SSCP approach is of additional value for assessment of the ongoing processes. The Enterobacteriaceae-specific PCR products could only be amplified from rhizosphere samples, which indicates a pronounced harboring or even active growth of these organisms in the rhizosphere. Differences in the community composition depending on the depth and the distance from inflow are shown in the SSCP profiles (Fig. 7). In the first 10 cm depth of the filters, the diversity was lower than in deeper parts; and the diversity decreased overall with the distance from inflow. Some bands (a and c; see Fig. 7) were present in all profiles while others disappeared (j, d, and e) or their intensity merely faded (h, i, k, and e) up to the outflow of the filter. Using the developed primers, we were able to show the distribution of the diversity of enteric bacteria in complex environmental systems for the first time. 3.3. Effect of operational conditions The enteric diversity of DNA extracted from rhizosphere samples varies with the depth and was higher in

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

1370

a

a

a

a

b c

b c

b c

b c

e

d

e f g

h i j k

1.5

d

e f g

h i j k

d

e f g

h i j k

2.5 3 .5 Distance from inflow [ m ]

30

20

10

30

20

10

30

20

10

30

20

10

Depth [ cm]

d f g

h i j k

4.5

Fig. 7. SSCP profiles of PCR-amplified (primer com1/ent18) partial 16S rRNA gene sequences of DNA extracted from the rhizosphere of a horizontal planted sand filter depending on the distance from inflow and the depth. Characters close to the bands show bands which (i) are abundant in most of the samples (bands a, b and c); (ii) appear with increasing depth (bands j and k) or (iii) disappear with increasing distance from the inflow (bands d, j and k).

the deeper parts of the horizontal filters (Fig. 7). The CFUs of heterotrophic bacteria from bulk soil decreased by up to one order of magnitude in the planted filter and by up to two orders of magnitude in the unplanted filter, depending on depth and distance from the inflow. The number of enteric bacteria in unplanted filters decreased to non-detectable levels (Fig. 4). It seems that most of the bacteria remained in the saturated zone of the unplanted filter at a depth of 40–60 cm, whereas in the planted filter the water was better distributed by the plant root as indicated by the higher volumetric water content at the different depths (data not shown). Powelson and Mills (2001) showed that the changes of the volumetric water content affect bacterial adsorption and that the bacterial transport by unsaturated flow is affected by adsorption to the air–water interfaces and to the solid phase. In vertical filters, a more or less homogeneous distribution of the bacteria over different depths was observed (Fig. 4) but specific bands were not present in certain profiles (data not shown). A homogeneous distribution of the bacteria with depth was recently shown for two constructed wetlands studied by DAPI staining of the water samples (Szewzyk et al., 2001).

3.4. Effect of filter materials The soil matrix seems to have a major effect on the enteric community in the rhizosphere of Phragmites australis although the same plant species was used (Fig. 8). Six to seven bands were observed in the SSCP profile of the rhizosphere from the sand, whereas more than 14 bands were detected in the rhizosphere of the plants growing in expanded clay. Such differences were also observed in the dendrogram where the profiles from the expanded clay and the sand clustered together (Fig. 6). The differences can be explained in relation to the soil material. Chiarini et al. (1998) showed that the rhizosphere of maize plants is markedly affected by the type and parameters of the soil. In the present wetland, the sand had a mixture of different grain sizes from 0.2 to 2 mm and a volumetric water content of 5–8% (wt/wt), whereas the expanded clay filter was composed of 25% sand with expanded clay particles having a grain size of 2–4 mm and a volumetric water content of 20–30% (wt/ wt) (data not shown). The salinity of the expanded clay plots was around 2700–5000 [mS/cm], whereas in the sand plots values of 700–1100 [mS/cm] were measured.

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373

35

15

45

35

15

Depth [cm]

1371

technology, filter material and plants each exert a strong selective influence on the microbial community in constructed wetlands. The overall efficiency of bacteria removal (499%) in constructed wetlands with different operational processes could be proven by use of the CFU methods. However, only the application of molecular profiling methods based on 16S rDNA was able to provide a deeper insight into the shifts of the microbial communities in the filter plots. Using the developed primers for Enterobacteriaceae combined with the PCR–SSCP method, we were able to determine which taxa were removed from the wastewater inflow and to show a harboring effect for Enterobacteriaceae in the rhizosphere. The applied methods can hence improve our understanding of the ‘black box’ of constructed wetlands. However, to trace certain pathogenic bacteria, more specific gene markers should be targeted.

Acknowledgments This work was kindly supported by the German Ministry of Education and Research (BMBF) (Grant 02WA0108). We would like to thank M. Alexandrino and U. Szewzyk (Berlin University of Technology, Germany) for providing the DNA of reference bacteria used for PCR optimization and Antonis Chatzinotas for helpful suggestions on the manuscript.

References

(A)

(B)

Fig. 8. SSCP profiles of PCR-amplified (primers Com1/Ent18) partial 16S rRNA gene sequences of DNA extracted from the rhizospheres of vertical sand (A) and expanded clay (B) filters.

These characteristics obviously have a dramatic effect on the harbouring and transport of bacteria and on biofilm formation in the expanded clay matrix.

4. Conclusions Although the same wastewater inflow was used for all the plots, the bacterial community was different in each unplanted filter, while the rhizosphere of Phragmites australis was colonized by distinct communities depending on the filter material. These differences show that the

Alexandrino, M., Grohmann, E., Szewzyk, U., 2004. Optimization of PCR-based methods for rapid detection of Campylobacter jejuni, Campylobacter coli and Yersinia enterocolitica serovar 0:3 in wastewater samples. Water Res. 38, 1340–1346. Altschul, S., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1998. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Faseb J. 12, A1326–A1326. Amann, R.I., Ludwig, W., Schleifer, K.H., 1995. Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol. Rev. 59, 143–169. APHA, 1998. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, DC. Arias, C.A., Cabello, A., Brix, H., Johansen, N.H., 2003. Removal of indicator bacteria from municipal wastewater in an experimental two stage vertical flow constructed wetland system. Water Sci. Technol. 48 (5), 35–41. Armstrong, W., Armstrong, J., Beckett, P.M., 1990. Measurement and modelling of oxigen release from roots of Phragmites australis. In: Cooper, F. (Ed.), Constructed Wetlands in Water Pollution Control. Pergamon Press, Cambridge.

ARTICLE IN PRESS 1372

G. Vacca et al. / Water Research 39 (2005) 1361–1373

Axelrood, P.E., Clarke, A.M., Radley, R., Zemcov, S.J.V., 1996. Douglas-fir root-associated microorganisms with inhibitory activity towards fungal plant pathogens and human bacterial pathogens. Can. J. Microbiol. 42, 690–700. Baeder-Bederski, O., Kuschk, P., Mosig, P., Mu¨ller, R.A., Borneff-Lipp, M., Du¨rr, M., 2004. Reducing faecal germs in municipal sewage using planted soil filters: initial results of a pilot plant system. Acta Horticult. 643, 257–263. Bassam, B.J., Caetano-Anolles, G., Gresshoff, P.M., 1991. Fast and sensitive silver staining of DNA in polyacrylamide gels. Anal. Biochem. 196, 80–83. Berg, G., Marten, P., Ballin, G., 1996. Stenotrophomonas maltophilia in the rhizosphere of oilseed rape—occurrence, characterization and interaction with phytopathogenic fungi. Microbiol. Res. 151, 19–27. Berg, G., Roskot, N., Steidle, A., Eberl, L., Zock, A., Smalla, K., 2002. Plant-dependent genotypic and phenotypic diversity of antagonistic rhizobacteria isolated from different Verticillium host plants. Appl. Environ. Microbiol. 68, 3328–3338. Bernard, J.M., Lauve, T.E., 1995. A comparison of growth and nutrient uptake in Phalaris arundinacea L. growing in a wetland and a constructed bed receiving landfill leachate. Wetlands 15, 176–182. Brix, H., 1997. Do macrophytes play a role in constructed treatment wetlands? Water Sci. Technol. 35 (5), 11–17. Bull, A.T., Ward, A.C., Goodfellow, M., 2000. Search and discovery strategies for biotechnology: the paradigm shift. Microbiol. Mol. Biol. Rev. 64, 573–606. Burger, G., Weise, G., 1984. Untersuchungen zum EinfluX limnischer Makrophyten auf die Absterbegeschwindigkeit von Escherichia coli im Wasser. Acta Hydrochim. Hydrobiol. 12, 301–309. Chiarini, L., Bevivino, A., Dalmastri, C., Nacamulli, C., Tabacchioni, S., 1998. Influence of plant development, cultivar and soil type on microbial colonization of maize roots. Appl. Soil Ecol. 8, 11–18. Decamp, O., Warren, A., 1998. Bacteriovory in ciliates isolated from constructed wetlands (reed beds) used for wastewater treatment. Water Res. 32, 1989–1996. Decamp, O., Warren, A., 2000. Investigation of E. coli removal in various designs of subsurface flow wetlands used for wastewater treatment. Ecol. Eng. 14, 293–299. Decamp, O., Warren, A., Sanchez, R., 1999. The role of ciliated protozoa in subsurface flow wetlands and their potential as bioindicators. Water Sci. Technol. 40 (3), 91–98. Dutka-Malen, S., Evers, S., Courvalin, P., 1995. Detection of glycopeptide resistance genotypes and identification to the species level of clinically relevant enterococci by PCR. J. Clin. Microbiol. 33, 1434. Friedrich, U., Van Langenhove, H., Altendorf, K., Lipski, A., 2003. Microbial community and physicochemical analysis of an industrial waste gas biofilter and design of 16S rRNAtargeting oligonucleotide probes. Environ. Microbiol. 5, 183–201. Gavini, F., Mergaert, J., Beji, A., Mielcarek, C., Izard, D., Kersters, K., De Ley, J., 1989. Transfer of Enterobacter agglomerans (Beijerinck, 1888) Ewing and Fife 1972 to Pantoea gen. nov. as Pantoea agglomerans comb. nov. and description of Pantoea dispersa sp. nov. Int. J. Syst. Bacteriol. 39, 337–345.

Gersberg, R.M., Gearhart, R.A., Yves, M., 1989a. Pathogen removal in constructed wetlands. In: Hammer, D.A. (Ed.), Constructed Wetlands for Wastewater Treatment; Municipal, Industrial and Agricultural. Lewis Publisher, Chelsea, Michigan, USA, pp. 431–446. Gersberg, R.M., Lyon, S.R., Brenner, R., Elkins, B.V., 1989b. Integrated wastewater treatment using artificial wetlands: a gravel marsh case study. In: Hammer, D.A. (Ed.), Constructed Wetlands for Wastewater Treatment; Municipal, Industrial and Agricultural. Lewis Publishers, Chelsea, Michigan, USA, pp. 145–152. Green, M.B., Griffin, P., Seabridge, J.K., Dhobie, D., 1997. Removal of bacteria in subsurface flow wetlands. Water Sci. Technol. 35 (5), 109–116. Hench, K.R., Bissonnette, G.K., Sexstone, A.J., Coleman, J.G., Garbutt, K., Skousen, J.G., 2003. Fate of physical, chemical, and microbial contaminants in domestic wastewater following treatment by small constructed wetlands. Water Res. 37, 921–927. Higgins, M.J., Rock, C.A., Bouchard, R., Wengrezynek, B., 1993. Controlling agricultural runoff by use of constructed wetlands. In: Moshiri, G.A. (Ed.), Constructed Wetlands for Water Quality Improvement. Lewis Publishers, Boca Raton, FL, pp. 359–367. Ibekwe, A.M., Grieve, C.M., Lyon, S.R., 2003. Characterization of microbial communities and composition in constructed dairy wetland wastewater effluent. Appl. Environ. Microbiol. 69, 5060–5069. Karanthanasis, A.D., Thompson, Y.L., 1995. Mineralogy of iron precipitates in a constructed acid mine drainage wetland. Soil Sci. Soc. Am. J. 59, 1774–1779. Kickuth, R., Kaitzis, G., 1975. Mikrobizid wirksame aromaten aus Scirpus lacustris L. Umweltschutz 4–5, 134–135. Lantz, P.G., Knutsson, R., Blixt, Y., Abu Al-Soud, W., Borch, E., Radstrom, P., 1998. Detection of pathogenic Yersinia enterocolitica in enrichment media and pork by a multiplex PCR: a study of sample preparation and PCR-inhibitory components. Int. J. Food Microbiol. 45, 93–105. Lottmann, J., Heuer, H., Smalla, K., Berg, G., 1999. Influence of transgenic T4-lysozyme-producing potato plants on potentially beneficial plant-associated bacteria. FEMS Microbiol. Ecol. 29, 365–377. Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar, Buchner, A., Lai, T., Steppi, S., Jobb, G., Forster, W., Brettske, I., Gerber, S., Ginhart, A.W., Gross, O., Grumann, S., Hermann, S., Jost, R., Konig, A., Liss, T., Lussmann, R., May, M., Nonhoff, B., Reichel, B., Stamatakis, A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K.H., 2004. ARB: a software environment for sequence data. Nucl. Acids Res. 32, 1363–1371. Maidack, B.L., Olsen, G.J., Larson, N., Overbeek, R., Mc Caughey, M.J., Woese, C.R., 1997. The RDP (Ribosomal Database Project). Nucl. Acids Res. 25, 109–111. Mashauri, D.A., Mulungu, D.M.M., Abdulhussein, B.S., 2000. Constructed wetland at the University of Dar Es Salaam. Water Res. 34, 1135–1144. Morales, A., Garland, J.L., Lim, D.V., 1996. Survival of potentially pathogenic human-associated bacteria in the rhizosphere of hydroponically grown wheat. FEMS Microbiology Ecology 20, 155–162.

ARTICLE IN PRESS G. Vacca et al. / Water Research 39 (2005) 1361–1373 Muyzer, G., Smalla, K., 1998. Application of denaturing gradient gel electrophoresis (DGGE) and temperature gradient gel electrophoresis (TGGE) in microbial ecology. Antonie Van Leeuwenhoek Int. J. Gen. Mol. Microbiol. 73, 127–141. Nikolausz, M., Ma´rialigeti, K., Kova´cs, G., 2004. Comparison of RNA- and DNA-based species diversity investigations in rhizoplane bacteriology with respect to chloroplast sequence exclusion. J. Microbiol. Methods 56, 365–373. Normander, B., Prosser, J.I., 2000. Bacterial origin and community composition in the barley phytosphere as a function of habitat and presowing conditions. Appl. Environ. Microbiol. 66, 4372–4377. Perkins, J., Hunter, C., 2000. Removal of enteric bacteria in a surface flow constructed wetland in Yorkshire, England. Water Res. 34, 1941–1947. Pierson, L.S., Pierson, E.A., 1996. Phenazine antibiotic production in Pseudomonas aureofaciens: role in rhizosphere ecology and pathogen suppression. FEMS Microbiol. Lett. 136, 101–108. Powelson, D.K., Mills, A.L., 2001. Transport of E. coli in sand columns with constant and changing water contents. J. Environ. Qual. 30, 238–245. Pru¨ss, A., Kay, D., Fewtrell, L., Bartram, J., 2002. Estimating the burden of disease from water, sanitation, and hygiene at a global level. Environ. Health Perspect. 110, 537–542. Pundsack, J., Axler, R., Hicks, R., Henneck, J., Nordmann, D., McCarthy, B., 2001. Seasonal pathogen removal by alternative on-site wastewater treatment systems. Water Environ. Res. 73, 204–212. Rivera, F., Warren, A., Ramirez, E., Decamp, O., Bonilla, P., Gallegos, E., Caldero´n, A., Sa´nchez, J.T., 1995. Removal of pathogens from wastewater by the root zone method (RZM). Water Sci. Technol. 32 (3), 211–218.

1373

Schwieger, F., Tebbe, C.C., 1998. A new approach to utilize PCR-single-strand-conformation polymorphism for 16S rRNA gene-based microbial community analysis. Appl. Environ. Microbiol. 64, 4870–4876. Sedlak, J., 1973. Present knowledge and aspects of Citrobacter. Curr. Top. Microbiol. Immunol. 62, 41–59. Smalla, K., Wieland, G., Buchner, A., Zock, A., Parzy, J., Kaiser, S., Roskot, N., Heuer, H., Berg, G., 2001. Bulk and rhizosphere soil bacterial communities studied by denaturing gradient gel electrophoresis: plant-dependent enrichment and seasonal shifts revealed. Appl. Environ. Microbiol. 67, 4742–4751. Soto, F., Garcı´ a, M., de Luı´ s, E., Be´cares, E., 1999. Role of Scirpus lacustris in bacterial and nutrient removal from wastewater. Water Sci. Technol. 40 (3), 241–247. Szewzyk, U., Alexandrino Fernandes, M.A., Grohmann, E., 2001. Untersuchungen zur mikrobiellen O¨kologie von Bodenfiltern als Voraussetzung zum Versta¨dnis der Elimination von Krankheitserregern. In: Fehr, G.C. (Ed.), Forschungsansa¨tze und Ergebnisse des Verbundprojektes Bewachsene Bodenfilter-DBU. DBU, pp. 73–80. Vrhovsek, D., Kukanja, V., Bulc, T., 1996. Constructed wetland (CW) for industrial waste water treatment. Water Res. 30, 2287–2292. Williams, J., Bahgat, M., May, E., Ford, M., Butler, J., 1995. Mineralisation and pathogen removal in gravel bed hydroponic constructed wetlands for wastewater treatment. Water Sci. Technol. 32 (1), 49–58. Yang, L., Chang, H.-T., Huang, M.-N.L., 2001. Nutrient removal in gravel- and soil-based wetland microcosms with and without vegetation. Ecol. Eng. 18, 91–105. Zablotowicz, R.M., Hoagland, R.E., Locke, M.A., Hickey, W.J., 1995. Glutathione-S-transferase activity and metabolism of glutathione conjugates by rhizosphere bacteria. Appl. Environ. Microbiol. 61, 1054–1060.