Natural organic matter (NOM) removal and structural changes in the bacterial community during artificial groundwater recharge with humic lake water

ARTICLE IN PRESS WAT E R R E S E A R C H

41 (2007) 2715 – 2725

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Natural organic matter (NOM) removal and structural changes in the bacterial community during artificial groundwater recharge with humic lake water Reija E. Kolehmainena,, Jo¨rg H. Langwaldtb, Jaakko A. Puhakkaa a

Institute of Environmental Engineering and Biotechnology, Tampere University of Technology, P.O. Box 541, FIN-33101 Tampere, Finland Geological Survey of Finland, Mineral Processing, Tutkijankatu 1, 83500 Outokumpu, Finland

b

art i cle info

ab st rac t

Article history:

This study evaluated the removal of natural organic matter (NOM) and structural changes

Received 30 June 2006

in the microbial community during infiltration of humic lake water at three artificial

Received in revised form

groundwater recharge (AGR) sites in Finland. The three sites were at waterworks in

19 December 2006

Ha¨meenlinna, Jyva¨skyla¨ and Tuusula, sites A, B and C, respectively. Site A used ground-

Accepted 22 February 2007

water recharge by both basin and sprinkling infiltration, site B used only sprinkling

Available online 16 April 2007

infiltration, and site C used only basin infiltration. Reductions of total organic carbon at

Keywords:

sites A, B and C were 91%, 84% and 74%, respectively, in the winter, and 88%, 77% and 73%,

Artificial groundwater recharge

respectively, in the summer. The Finnish national recommended value of 2 mg/l for TOC

Bacterial community structure

was achieved at all sites and the TOC of natural groundwater at site C was much lower, at

Basin infiltration

0.6 mg/l. Large molecular fractions of NOM were removed more efficiently than the smaller

DGGE

ones. Total amount of DAPI-stained cells decreased during infiltration at sites A, B and C in

NOM

winter by 94%, 94% and 75% and in summer by 96%, 97% and 94%, respectively. Bacterial

16S rRNA gene

communities in raw waters and extracted groundwaters were diverse with changes

Sprinkling infiltration

occurring during infiltration, which was shown by DNA extraction followed by PCR of 16S rRNA genes and denaturing gradient gel electrophoresis (DGGE) fingerprinting. While the natural groundwater microbial community was diverse, it was different from that of the extracted groundwater in the AGR area. Simultaneous organic carbon removal and the decrease of bacterial counts during infiltration indicated biodegradation. In addition, the changing DGGE profiles during the process of infiltration, demonstrated that changing environmental conditions were reflected by changes in bacterial community composition. & 2007 Elsevier Ltd. All rights reserved.

1.

Introduction

One of the major aims of drinking water treatment is to remove natural organic matter (NOM) from raw water. NOM may react with disinfection chemicals and cause undesirable microbial growth in distribution systems (Pre´vost et al., 1998). During artificial groundwater recharge (AGR), NOM is removed by physical, chemical and microbial Corresponding author. Tel.: +358 3 3115 2851; fax: +358 3 3115 2869.

E-mail address: [email protected] (R.E. Kolehmainen). 0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2007.02.042

processes. In conventional basin infiltration, part of the removal occurs in sediment biofilms of the basin (Kivima¨ki, 2001). In sprinkling infiltration, forest soil also releases NOM to the subsurface (Lindroos et al., 2002). NOM is then removed in the vadose and saturated zones. While AGR is a relatively old and widely used method of removing NOM, little is known about its microbial ecology and the role of microorganisms in NOM removal.

ARTICLE IN PRESS 2716

WAT E R R E S E A R C H

41 (2007) 2715– 2725

Transport of large quantities of NOM into aquifers is a crucial issue of AGR due to the potential risks of clogging (Pe´rez-Paricio and Carrera, 2001 and references therein) and break-through. Either of these would adversely affect aquifers. Microorganisms may play a key role in preventing these adverse effects and providing a method of environmentally sustainable drinking water production. Several studies have focused on the different NOM removal mechanisms in both AGR and soil aquifer treatment (SAT) (e.g. Kortelainen and Karhu, 2006; Rauch-Williams and Drewes, 2006; Rauch and Drewes, 2005; La˚ngmark et al., 2004; Juhna et al., 2003; Hendel et al., 2001; Kivima¨ki, 2001; Drewes and Fox, 1999; Drewes and Jekel, 1998; Miettinen, et al., 1996) illustrating the importance of biodegradation. Bacterial abundance in the subsurface varies in different zones depending on the hydrological, physical, and geochemical conditions. In the vadose zone, the decline of nutrients results in a decline in bacterial numbers. However, possibly due to mixing of oxygen and recently recharged nutrients, the groundwater interface has a higher number of bacteria (Madsen and Ghiorse, 1993). In the subsurface, most of the bacteria are attached to soil particles (Hazen et al., 1991) as biofilms. These biofilms receive nutrients from flowing water (van der Kooij et al., 1995) and create potential sorption sites for dissolved substances by producing hydrophilic extracellular polymeric substances (EPS) (Flemming, 1995). Hazen et al. (1991) found less than 1  106–5  108 bacteria/gdw of subsurface sediment and 1  103–6.3  104 bacteria/ml of adjacent groundwater by direct counting. The communities of attached and unattached bacteria are different but include overlapping components of a dynamic community (Madsen and Ghiorse, 1993). The majority of attached microorganisms in pristine groundwater communities are likely to also have motile, unattached stages (Hirsch and Rades-Rohkohl, 1983). The rate and extent of exchange between attached and unattached bacteria is likely to be regulated by the availability of electron donors and acceptors, inorganic and organic nutrients, as well as other factors that impact growth, attachment and life cycles of subsurface bacteria (Madsen and Ghiorse, 1993). The NOM content may be a significant factor in controlling microbial and other processes in the subsurface (Aiken, 2002). However, microorganisms have a remarkable capacity to adapt to changing environments by natural selection, mutation, and nongenetic adaptation (Alexander, 1965). Also, horizontal gene transfer is a common process in natural and engineered environments (Wuertz et al., 2004; Tiirola et al., 2002). NOM is not easily biodegradable due to its

complex polymeric structure with aromatic units and many types of covalent bonds and cross-linking within the organic macromolecular structure (Marschner and Kalbitz, 2003; Kalbitz et al., 2003; Ka¨stner and Hofrichter, 2001; Thurman, 1985; Alexander, 1965). This complex macromolecular structure requires diverse groups of microorganisms for stepwise biodegradation. Little is currently known about the microorganisms that transform and recycle NOM in aquifers used for AGR. This study aims to: (i) evaluate removal of NOM and its different molecular size fractions during AGR, (ii) characterise structure and dynamics of microbial communities, (iii) compare extracted artificially recharged groundwater quality with natural groundwater quality and chemically treated water quality, (iv) assess whether raw water quality changes in sprinkling networks and an infiltration basin, (v) examine seasonal trends and finally, (vi) study correlations between organic carbon removal and cell counts during infiltration. Due to inaccessibility of sediment samples only water samples were studied.

2.

Materials and methods

2.1.

Study sites and groundwater sampling

The recharge study sites were unconfined aquifer glacial formations, located in public water works of Ha¨meenlinna, Jyva¨skyla¨ and Tuusula, in Finland. These three sites are referred to as sites A, B and C, respectively. Selected site parameters are shown in Table 1. The first set of samples was collected in winter (January 2005), when lakes were covered in ice and the water temperature was low (approximately 5 1C). The second set of samples was collected in summer (August and September 2005), when lake water temperature was moderate (approximately 20 1C). Samples at each site were taken from: (1) the raw water, (2) the monitoring well (approximately half-way between the raw water infiltration site and the extraction well) and (3) the extracted groundwater (Fig. 1). Additionally, in the summer, samples were taken from an infiltration basin (site C) and sprinkling networks (sites A and B). These were compared to the raw water samples to evaluate whether there were any differences. Also, at site C, extracted groundwater was compared to natural groundwater (270 m from the infiltration basin) that was not under the influence of surface water. The quality of extracted groundwater at AGR sites was compared with chemically treated potable water from the Tampere water works (site D), where the lake water purification process

Table 1 – Characteristics of groundwater recharge sites Symbol

Method

Start of recharge

Recharge (m3/d)

Infiltration distance (m)

Estimated retention time (d)

Ha¨meenlinna

A

1000–1300

90

B C

1976 1995 2000 1979

20,000

Jyva¨skyla¨ Tuusula

Basin Sprinkling Sprinkling Basin

15,000 10,000

200–550 500–700

15–30 30–60

Site

ARTICLE IN PRESS WAT E R R E S E A R C H

Nat.

Table 2 – Equations used for calculating aromaticity values for the water samples

S/B MW Ext.

Raw

Aromaticity % ¼ 527  SUVA254+2.8 SUVA254 ¼ UVA254/DOC Aromaticity % ¼ 0.057  e280+3.001 e280 ¼ UVA280  12,000/(l  DOC) Nomenclature SUVA254 UVA254

A

3500m

700-1000m

300m

B

7300m

150-500m

50m

C

>12000m

200-400m

300m

Fig. 1 – Schematic diagram of sampling locations at artificial groundwater recharge sites. Abbreviations used in the text: Raw, raw water; S/B, sprinkling network or basin water; MW, monitoring well water; Ext., extracted groundwater; Nat., natural groundwater; Chem. (not in figure), chemically treated water from the Tampere water works.

includes addition of CO2 and quicklime, chemical precipitation with Fe2(SO4)3, flotation, addition of ClO2, rapid sand filtration, activated carbon filtration and addition of Cl2. Prior to sampling, stagnant water was removed from the monitoring wells by pumping. Samples were taken from 1 m below groundwater table. Microbial samples were stored at 5 1C in the dark until they were processed. Chemical samples were acidified on site with HCl to reach pH below 2 and stored at 5 or 20 1C.

2.2.

2717

4 1 (200 7) 271 5 – 272 5

Chemical analysis and microbial counts

Samples were acidified and filtered through 0.45 mm cartridge filters (Polyethersulfone Membrane, Whatman) for measurements of absorbance at 254 and 280 nm, DOC, and the organic matter molecular size distribution. The molecular size distribution was measured by high performance liquid chromatography (HPLC) (Hewlett Packard 1100 series) at 254 nm with a size exclusion chromatography (SEC) column (TSK-GEL G3000W) using sodium acetate as the eluent. For the initial approximation of the molecular sizes, the calibration curve for the HPSEC was made by using polystyrene sulphonate standards (2200, 3600 Da by Polymer Standards Service (PSS) GmbH, Krotek Ltd. and 210, 4300, 6800, 13000 Da by Fluka) and acetone (58 Da). The calibration curve of log MW versus retention time was described by Eq. (1): y ¼ 0.3018x+5.2781 (R2 ¼ 0.95). Filters for the analyses were pre-rinsed with 200 ml of deionised water. UV-absorbance at 254 and 280 nm (Shimadzu pharmaspec UV-1700) and DOC/TOC (Shimadzu TOC-5000A) measurements were performed at Tampere water works, Finland. Relative dissolved aromatic carbon contents were estimated using Eqs. (2) and (3) for winter and summer samples. For comparison, Eqs. (4) and (5) were also used for summer samples (Table 2). Total cell counts in water samples were performed within 24 h of sample collection by filtering 2–20 ml of water through 0.2 mm membrane filters (Millipore) and staining the cells with DAPI (40 ,6-diamidino-2-phenylindole)

DOC e280 UVA280 1

(2) (3) (4) (Peuravuori and Pihlaja, 1997) (5) Specific UV absorbance at 254 nm (1/mg cm) UV light absorbance at 254 nm Dissolved organic carbon concentration (mg/l) Specific molar absorbance at 280 nm (1/Mol l cm) UV light absorbance at 280 nm Cell length (cm)

nucleic acid staining solution (Molecular Probes). Cells were counted 20 times, and processed in duplicate using an epifluorescense microscope (ZEISS Axioscop 2). Site D TOC values were obtained from the water works and represent the average TOC value from 2005 (n ¼ 236 for raw water and n ¼ 48 for chemically purified water).

2.3.

Molecular biology analysis

Three hundred to 4200 ml of sample was filtered for DNA extraction depending on the solid content of the sample. Filters were stored at 20 1C until DNA extraction by physical and chemical lyses of the cells. In winter, duplicate water samples were filtered using two different filters: 0.2 mm pore size hydrophilic polycarbonate filter (Whatman) and sterile 0.2 mm pore size glass fibre filter membrane (Schleicher & Schuel). As DNA yields were similar, only the former were used in summer. Filters were inserted to tubes and the extraction was performed according to the kit protocol (Ultraclean Soil DNA kit, Mobio Laboratories Inc.). The crude DNA sample was used as a template for polymerase chain reaction (PCR). A pair of Eubacterial primers GC-BacV3f (50 CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCA CGG GGG GCC TAC GGG AGG CAG CAG-30 , including GC clump for DGGE) and reverse 907r (50 -CCG TCA ATT CMT TTG AGT TT-30 ) were used to amplify a 16S rRNA gene fragment with a length of approximately 570 bp. The PCR master mix contained 5 ml of 10  buffer (1  is 10 mM Tris-HCl, pH 8.8 at 25 1C, 1.5 mM MgCl2, 50 mM KCl and 0.1% Triton X-100), 0.5 mM of each primer, 100 mM of each deoxynucleotide triphosphate, 1.5 U of DyNAzyme II DNA polymerase, 0.02 mg of bovine serum albumin, and sterile DNase RNase free water to give a final volume of 49 ml. Forty-five to 48 ml of master mix was used with 5–2 ml of template in the PCR, depending on the DNA yield. PCR was performed to amplify the DNA in a thermal cycler (Peltier Thermal Cycler PTC-200 MJ Research and Biometra T3000) with the following program: initial denaturation at 95 1C for 5 min, 30 cycles of 94 1C for 30 s, 50 1C for 1 min, 72 1C for 2 min, and final extension at 72 1C for 10 min.

ARTICLE IN PRESS 2718

WAT E R R E S E A R C H

41 (2007) 2715– 2725

The amount of product was visualised in 1% (w/v) agarose gel electrophoresis using ethidium bromide as a stain. Denaturing gradient gel electrophoresis (DGGE) was performed with a universal mutation detection system (BIO-RAD Hercules CA in winter, INGENYphorU2 Ingeny International BV the Netherlands in summer). Samples were loaded into 8% (w/v) polyacrylamide gels (bisacrylamide gel stock solution) in 1  TAE (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8.3) with denaturing gradients ranging from 30% to 65% (100% denaturant contains 7 M urea and 40% (v/v) formamide). Electrophoresis was run at 60 1C with 100 V for 16 h (BIO-RAD) or 22 h (INGENY phorU2). After the run, the gel was stained using SYBR Gold solution (Invitrogen, Molecular Probes) and photographed on a UV transilluminator with a Kodak DC290 digital camera. Selected DNA fragments were cut from the gel and eluted in 20 ml sterile water over night. Eluate (4 ml) was used as a template in PCR with primers BACV3f (50 -CCT ACG GGA GGC AGC AG-30 ) and 907r using PCR program described above. Purification and sequencing were performed at the DNA Sequencing Facility, Institute of Biotechnology, Helsinki University. Sequences were compared to those in databases using BLAST (http://www.ncbi.nlm.nih.gov/BLAST/). Genetic similarities between different samples were compared by cluster analysis for DGGE patterns with GelCompar II 3.0 software (Sint-Martens-Latem).

2.4.

Nucleotide sequence accession numbers

16 14 12 10 8 6 4 2 0

January TOC [mg/l]

TOC [mg/l]

The sequenced DGGE bands obtained in this study have GenBank accession numbers DQ784578-DQ784605.

Raw

MW

Ext.

Site A

Site B

Site C

0.6

0.3 0.2

3.1.

Organic carbon removal

Organic carbon removal and changes in molecular weight distribution of NOM during infiltration were studied. Absorbance and TOC results were as shown in Fig. 2. DOC concentrations comprised on average 88% and 93% of the TOC concentrations in all samples in winter and summer, respectively. At site A, initial contents of DOC (not shown) and TOC of raw water were higher than at sites B and C during both seasons. However, the lowest DOC and TOC concentrations were also observed in extracted groundwater from site A. Raw water at site C had the lowest concentrations of DOC and TOC in winter 2005, but the removal efficiency was also lowest. The lake water TOC increased slightly in sites A and B sprinkling networks, and decreased in the site C basin. Natural groundwater TOC was 37% lower than extracted groundwater from site C. Treated water from all AGR sites had lower TOC values than site D. HPSEC resulted in 5–6 NOM fractions (I–VI) in raw waters. Based on the calibration curve, estimations of molecular sizes [Da] for different fractions were: I: 2000–3300, II: 1800–2200, III: 1400–2000, IV: 1000–1400, V: 500–1000, VI: 250–500. The molecular size distribution of NOM changed significantly during infiltration at all sites (Fig. 3(A–C) as examples of summer samples). Total height of the HPSEC peaks (Fig. 3(D–E)), as well as the peak area (not shown), decreased during infiltration. This indicates a gradual decrease of total

16 14 12 10 8 6 4 2 0

August-September

S/B Site A

MW Site B

Ext.

Nat.

Site C

Site D

0.30

Chem.

August-September

0.25 Abs. 254

Abs. 254

0.4

Results

Raw

January

0.5

3.

0.20 0.15 0.10 0.05

0.1

0.00

0.0 Raw Site A

MW Site B

Ext. Site C

Raw

S/B Site A

MW Site B

Ext.

Nat.

Site C

Fig. 2 – Total organic carbon removal during infiltration measured as TOC (A–B) and UVA254 (C–D) at water works A, B and C in winter (A and C) and summer (B and D) 2005. In (B), values for raw and chemically treated water from site D are presented. Abbreviations are explained in Fig. 1 and Table 1. Error bars represent the standard deviation of two replicates.

ARTICLE IN PRESS

Peak Height [mAU]

WAT E R R E S E A R C H

16 14 12 10 8 6 4 2 0

Site A September

Raw

S

MW

II III IV V

VI

3.0 Peak Height [mAU]

Ext.

Site B August

2.5 2.0 1.5 1.0 0.5 0.0

Raw

S

MW

II III IV V

Ext.

VI

Peak Height [mAU]

3.0

Site C September

2.5 2.0 1.0

NOM in water during infiltration. At site B’s sprinkling network and site C’s basin, the total HPSEC peak height increased slightly compared to raw water, whereas at site A’s sprinkling network, the peak height sum decreased sharply. Large molecular fractions decreased more than the smaller ones at all sites. The largest organic matter fraction, fraction I, occurred only in raw water from site A in the winter (not shown). Even though the highest sum of HPSEC peak heights was found in raw water samples from site A in winter, the extracted groundwater had the lowest sum of the peak heights at this site, showing the greatest removal. In natural groundwater at site C, fewer peaks (fractions IV–VI) with lower peak heights occurred than in the extracted groundwater. Comparison of the different formulas for estimation of relative dissolved aromatic carbon contents resulted in a strong Pearson’s correlation; r-values were 0.98, 0.96 and 0.84 for sites A, B and C, respectively. Relative aromaticity (based on Eqs. (2) and (3)) increased slightly in the sprinkling networks and the basin (Fig. 4(B)). Apart from site B, relative aromatic carbon content reached the lowest value in the final sampling point (extracted groundwater), in both seasons. Relative aromaticity of natural groundwater in the summer at site C (6.3%) was slightly less than in the extracted groundwater (9.6%).

3.2.

0.5 B

Raw

MW

Ext.

II III IV V

Sum of Peak Heights [mAU]

2719

1.5

0.0

35 30 25 20 15 10 5 0

VI

January

Raw Site A

Sum of Peak Heights [mAU]

4 1 (200 7) 271 5 – 272 5

MW

Ext.

Site B Site C

45 40 35 30 25 20 15 10 5 0

August-September

Raw

B Site A

MW

Ext.

Site B Site C

Nat.

NAT.

Microbial community structure and dynamics

Changes in total cell counts, bacterial community diversity, and dynamics in the subsurface during infiltration were also studied. Total cell counts in raw waters were higher in the summer than in the winter (Fig. 5). On the contrary, in extracted groundwater, apart from site A, total cell counts were lower in summer than in winter. Cell numbers decreased in the sprinkling networks and in the basin by 13.5%, 1.8% and 13.0%, when compared with the raw water at sites A, B and C, respectively. Despite the difference between cell numbers in raw water at sites A, B and C (24.0  105, 17.1  105 and 14.3  105 cells/ml, respectively), almost the same number of cells existed in extracted groundwater (1.0  105, 0.6  105 and 0.8  105 cells/ml, respectively) from all three sites in summer. Cell counts in natural groundwater at site C (in summer) were 0.5  105/ml. At site D in March 2006, chemically treated water contained 0.1  105 cells/ml which was even less than in site C natural groundwater. According to DGGE analyses, all raw waters and extracted groundwater samples contained diverse bacterial communities; the number of identifiable bands was 30–40 for all the samples (Figs. 6–9). Samples from site C’s natural groundwater showed 20–30 identifiable bands. Clear changes within

Fig. 3 – Changes in the concentrations (peak heights) of molecular weight distribution of NOM during infiltration at sites A, B and C in summer (A–C). Estimations of fraction sizes (Da): I, 2000–3300; II, 1800–2200; III, 1400–2000; IV, 1000–1400; V, 500–1000; VI, 250–500. Peak height sums at sites A, B and C in winter (D) and summer (E). Note the different scales on the y-axes.

ARTICLE IN PRESS 2720

WAT E R R E S E A R C H

41 (2007) 2715– 2725

20

25

August-September

20

Aromaticity [%]

Aromaticity [%]

January 15 10 5 0

Raw Site A

MW

10 5 0

Ext.

Site B

15

Raw

S/B

Site C

Site A

MW Site B

Ext.

Nat.

Site C

Fig. 4 – Changes in dissolved aromatic carbon content during infiltration at sites A, B and C in winter 2005 (A) and summer (B) using Eqs. (1) and (2).

30

30

August-September

January

25 Bacteria x 105 / ml

Bacteria x 105 / ml

25 20 15 10

20 15 10 5

5

0

0 Raw Site A

MW Site B

Ext.

Raw

S/B

Site C

MW

Site A

Site B

Ext. Site C

Nat.

Chem.

Site D

Fig. 5 – Total microscopic cell counts of DAPI stained cells in water samples. Error bars indicate standard deviation of two replicates.

the community structure occurred during the recharge process. The clustering of DGGE band patterns was used to construct dendrograms as an illustration of the degree of genetic similarity between samples (Figs. 7 and 9). As a group, raw water samples had more similar band patterns than any other group of samples (S/B, MW, Ext.). At site B, the microbial community in monitoring well water and extracted groundwater samples clustered more closely together on the dendrogram (97% similar in summer and 83% similar in winter), compared to the other sites. Similarity between the microbial communities in natural groundwater and extracted groundwater at site C was only 35%. Microbial community structure did not change significantly in sprinkling networks or in the basin. Sequencing of the bands (winter samples) succeeded only for a few samples due to co-migration of several DNA fragments. Most of the bands that could be sequenced had closest relatives in uncultured bacterial clones. The majority of the species in raw waters had a close similarity with species originating from fresh waters. On the other hand, most of the species in monitoring well water at site B had the closest similarity with soil species. Characterised phylogenetic groups included Actinobacteria, Acidobacteria, Cyanobacteria and two Gram-positive bacteria (Table 3).

4.

Discussion

This study demonstrated efficient removal of NOM, changes within the bacterial community structure, and community dynamics during infiltration at three Finnish AGR sites in two seasons. The recommended TOC value of 2.0 mg/l for drinking water in Finland (Finnish Ministry of Social Affairs and Health, 1994) was reached at all sites (1.2–2.0 mg/l). The greatest NOM reductions occurred between the infiltration sites and the monitoring wells. DOC comprised on average 90% of TOC, which agrees with previous studies (e.g. Thurman, 1985). Helmisaari et al. (2003) found that retention time played a greater role in TOC removal than recharge distance. While this study does not agree with Helmisaari et al. (2003) findings, the number of the study sites was small and this difference could have been observed by chance. At site A, where both the recharge distance and the retention time were the greatest, the most efficient DOC removal occurred. Strong correlation existed between TOC and UVA254 values (r ¼ 0.9970.01), in agreement with a study by Wang and Hsieh (2001). HPSEC showed 5–6 NOM fractions (I–VI) in raw waters. Fraction I (2000–3300 Da) occurred only at site A in winter. A

ARTICLE IN PRESS WAT E R R E S E A R C H

4 1 (200 7) 271 5 – 272 5

Fig. 6 – DGGE fingerprints of amplified DNA fragments coding for Eubacterial 16S rRNA genes from the microbial populations in sites A, B and C in winter. Duplicate samples were used. Site A monitoring well water and extracted groundwater and site B monitoring well water were duplicated only from the point of DGGE due to the failure in DNA extraction of the true duplicate samples. Lanes std represent standard reference patterns.

100

90

80

70

60

50

Dice (Tol 0.5%-0.5%) (H>0.0% S>0.0%) [0.0%-100.0%]

Site B Ext. Site B MW Site C Ext. Site A Ext. Site A MW Site A Raw Site B Raw Site C Raw Site C MW Fig. 7 – Dendrogram illustrating the degree of genetic similarity of the communities estimated by clustering the DGGE patterns at sites A, B and C in winter (GelCompar II software).

2721

relatively strong correlation existed between TOC and the sum of peak heights (r ¼ 0.9370.05), however, small molecules are often invisible to the detector used in HPSEC (O’Loughlin and Chin, 2001) and were therefore eliminated from the analysis. Large molecular fractions were removed more efficiently than smaller ones, in agreement with findings reported by Helmisaari et al. (2003), Lindroos et al. (2002) and Kivima¨ki (2001). Removal of large fractions did not result in accumulation of smaller fractions. Higher quality water in terms of TOC was obtained at all AGR sites, compared to the chemically treated water at site D. Similarly, Lehtola et al. (2002) demonstrated that the mean TOC content of chemically purified water before disinfection at nine Finnish drinking water treatment plants was 2.2 mg/l (1.2–3.3 mg/l). The effect of dilution of recharged groundwater with natural groundwater was not considered in this study. However, in a previous study, Kortelainen and Karhu (2006) showed that site C’s dilution decreased DOC content during the final stage of infiltration (450–700 m) by 14%. Moreover, extracted groundwater was composed of 52% natural groundwater and 48% infiltrated lake water. In the DOC measurements, filters released some carbon, which shifted the results toward higher values. This was also seen in absorbance measurements. This is a general problem when dealing with low organic carbon concentrations (e.g. Karanfil et al., 2005). Raw water quality changed only slightly in the sprinkling networks and the basin, shown by slight TOC increases in the sprinkling networks and decreases in the basin. Minor reductions in total cell counts were observed in both systems. The extracted groundwater quality was similar in the sprinkling and basin infiltration sites in this study and in the study by Helmisaari et al. (2003). Even though sites A and C have been operating since the end of 1970s, and site B has only been operating since the end of 2000, organic matter removal efficiencies, cell counts, and the diversity of the bacterial community were similar during the recharge processes at all three sites. Relative dissolved aromatic carbon content estimates, based on both UVA254 and UVA280, were highly correlated (r ¼ 0.9370.04). The relative aromaticity increased slightly in the sprinkling networks and in the basin indicating preferential removal of non-aromatic compounds. This trend of preferential removal of non-aromatic compounds, however, did not continue during infiltration. On the contrary, the results indicated an efficient removal of aromatic compounds. This study demonstrated a strong positive correlation between organic carbon removal and total cell counts (r ¼ 0.9870.01). Cell counts decreased substantially during the recharge process at all sites, most likely due to adsorption (Yavuz Corapcioglu and Haridas, 1984 and references therein). Cell counts in raw waters were slightly higher in the summer than in the winter. Nevertheless, relative reductions of cell counts were greater in the summer, resulting in lower cell numbers in extracted groundwater at sites B and C and about the same number at site A. More efficient removal of cells in the summer may indicate more favourable sorption conditions. Extracted groundwater in the winter and summer samples contained an average 1.0  105, 0.7  105 and 1.3  105 cells/ml at sites A, B and C, respectively. Lehtola

ARTICLE IN PRESS 2722

WAT E R R E S E A R C H

41 (2007) 2715– 2725

Fig. 8 – DGGE fingerprints of amplified DNA fragments coding for Eubacterial 16S rRNA genes from the microbial populations in sites A, B and C in summer. Duplicate samples were used. Lanes std represent standard reference patterns.

100

80

60

40

20

Dice (Tol 0.5%-0.5%) (H>0.0% S>0.0%) [0.0%-100.0%]

Site C Basin Site C Raw Site A Raw Site B Raw Site B Spr. Site A Ext. Site A MW Site B Ext. Site B MW Site C Ext. Site C Nat. Site C MW Fig. 9 – Dendrogram illustrating the degree of genetic similarity of the communities estimated by clustering the DGGE patterns at sites A, B and C in summer (GelCompar II software).

et al. (2002) found that total bacterial numbers after chemical treatment (before final disinfection) at nine Finnish surface water works varied between 1.6  103 and 3.1  105 cells/ml. This indicates that bacterial numbers in extracted groundwater were higher than in some chemically purified water systems. At site D, chemical treatment resulted in extremely low cell numbers, 1.3  104 cells/ml. Natural groundwater (site C) had lower TOC and cell counts than the product of AGR. This is expected because a large quantity of NOM has been introduced to the subsurface during AGR. Despite the decrease in total cell counts and amounts of organic matter in water, the DGGE band patterns revealed diverse bacterial communities during the infiltration process. Both visual evaluation and cluster analysis showed changes within the bacterial community structures during infiltration. Natural groundwater had a diverse microbial community with one dominant species, i.e., one band with strong intensity. All studied raw waters showed similar band patterns. Comparison of raw water bacterial community structures with their corresponding sprinkling and basin water communities illustrated only minor variations in band intensities, i.e., the species dominance. Due to different volumes of water used for DNA extraction and different volumes of templates used for PCR, comparison of band intensities between samples is not appropriate. Microbial community profiles in extracted groundwater were clearly different from the natural groundwater community profiles,

ARTICLE IN PRESS WAT E R R E S E A R C H

2723

4 1 (200 7) 271 5 – 272 5

Table 3 – Selected DGGE band identities of AGR sites in winter samples No.

1 2 3 4 5 15 18 20 26 27 28 29 31 33 34 35 37 38 41 43 48 58 61 63 67 68 69 71

Sample

Similarity % (identities)

Phylogenetically closest organism in GenBank database (accession no.)

Phylogenetic group

Origin of the sample with the closest match

Site A raw Site A raw Site A raw Site A raw Site A raw Site A raw Site A MW Site A MW Site B raw Site B raw Site B raw Site B raw Site B raw Site B raw Site B raw Site B MW Site B MW Site B MW Site B MW Site B MW Site B ext Site B raw Site C raw Site C raw Site C raw Site C raw Site C MW Site C raw

99 (481/484)

Unknown Actinobacteria

Freshwater reservoir, metal-rich particles Freshwater habitat

Actinobacteria

Environmental sample

Unknown

Ultra-oligotrophic Crater lake

96 (417/431)

Uncultured bacterium clone HTC12 (AF418950) Uncultured actinobacterium clone NM2 (AJ575535) Uncultured Actinobacteria bacterium clone Act-80 (AY802927) Uncultured Crater Lake bacterium CL500-95 (AF316665) Gram-positive bacterium Z34 (AF488669)

Freshwater bacterioplankton

100 (430/430)

Uncultured lake bacterium P38.43 (AY752103)

Gram-positive bacterium Unknown

90 (192/211)

Uncultured bacterium clone GZKB19 (AJ853514) Uncultured cyanobacterium clone ST01-SN2C (AY222299) Uncultured actinobacterium clone NM2 (AJ575535) Uncultured Crater Lake bacterium CL500-95 (AF316665) Gram-positive bacterium Z34 16S ribosomal RNA gene Uncultured lake bacterium P38.2 (AY752084)

98 (412/420) 92 (443/481) 94 (452/478)

87 (359/412) 99 (479/480) 85 (366/430) 99 (427/429) 99 (457/460) 98 (470/478) 97 (469/481) 100 (471/471) 98 (336/341) 97 (330/341) 96 (423/440) 94 (395/420) 96 (424/440) 93 (382/410) 98 (443/451) 100 (447/447) 98 (475/480) 99 (488/491) 100 (466/466) 96 (452/470) 98 (463/470)

Uncultured bacterium clone 163ds20 (AY212615) Uncultured bacterium clone TLM09/ TLMdgge12a (AF534433) Uncultured lake bacterium P38.43 (AY752103) Uncultured bacterium clone EA1G3 (AY186069) Uncultured soil bacterium clone L1A.9C09 (AY989183) Uncultured bacterium clone 21BSF61 (AJ863264) Uncultured Acidobacteria bacterium clone EB1088 (AY395407) Uncultured bacterium clone: BS118 (AB240262) Uncultured bacterium clone uEV818BHEB5102702SAS79 (DQ256360) Uncultured bacterium clone TLM09/ TLMdgge12a (AF534433) Uncultured actinobacterium clone NM2 (AJ575535) Uncultured firmicute ESR 12 (AF268296)

Cyanobacteria

Oligomesotrophic lake bacterioplankton Leachate of a closed municipal solid waste landfill Environmental sample

Actinobacteria

Freshwater habitat

Unknown

Ultra-oligotrophic Crater lake

Unknown

Freshwater bacterioplankton

Unknown

Oligomesotrophic lake bacterioplankton Stream close to an equine manure pile Arctic lake bacterioplankton

Unknown

Unknown Unknown Unknown

Unknown

Oligomesotrophic lake bacterioplankton Ferromanganese deposits in Lechuguilla and Spider Caves Soil bacteria

Unknown

Soil bacteria

Acidobacteria

Pasture soil bacteria

Unknown

Soil rhizosphere

Unknown Unknown

Subsurface water of the Kalahari Shield, South Africa Arctic lake bacterioplankton

Actinobacteria

Freshwater habitat

Unknown

Freshwater bacteria during phytoplankton bloom Stream close to an equine manure pile Oligomesotrophic lake bacterioplankton Stream close to an equine manure pile Arctic lake

Unknown

Uncultured bacterium clone 163ds20 (AY212615) Uncultured lake bacterium P38.43 (AY752103)

Unknown

Uncultured bacterium clone 163ds20 (AY212615) Uncultured bacterium clone TLM09/ TLMdgge12a (AF534433)

Unknown

indicating that AGR changed the natural subsurface community structure. Kilb et al. (1998) reported similar results from comparisons of DGGE band patterns of water and sediment

Unknown

Unknown

samples in aerobic and anaerobic groundwater habitats during infiltration in Germany. They found great differences between aerobic and anaerobic groundwater and sediment

ARTICLE IN PRESS 2724

WAT E R R E S E A R C H

41 (2007) 2715– 2725

communities and significant changes during underground passage. Only water samples were used in the community analyses. Changes in microbial communities in relationship to NOM removal in the water phase during infiltration, however, partially reflects the impact of AGR on subsurface microbiology, as well as the role of biodegradation in water purification. The latter conclusion was confirmed by Kortelainen and Karhu (2006). They measured oxidative decomposition of DOC, as d13CDIC, at site C. This was the only process that consumed DOC and produced DIC within 330 m of infiltration. The oxidative decomposition decreased thereafter due to preferential removal by adsorption. Even though clear seasonal trends were not seen in NOM removal efficiency or the microbial community structure in this study, as Kortelainen and Karhu (2006) suggested, the proportion of biodegradation relative to adsorption may vary seasonally. It can be assumed that the free-living bacterial community reflects the soil-attached community, both of which are important in NOM biodegradation. According to Alfreider et al. (1997), the subsurface pore habitat is predominantly populated by free-living bacteria, but may exhibit much less community diversity and activity than attached bacteria in the sediment. Harvey and Barber (1992) and Bengtsson (1989) concluded that an increase in organic carbon level in an aquifer may promote the growth of free-living bacteria or stimulate the activity of attached bacteria that have unattached life cycle stages (Anderson and Lovley, 1997). In this study, the strong positive correlation between TOC removal and total cell counts, and the great diversity of bacterial species throughout infiltration in AGR, are indications of the important role of biodegradation in the purification process.

5.

Conclusions

The removal of NOM and structural changes in the bacterial community during infiltration with humic lake water at three Finnish AGR sites were investigated. The following conclusions can be drawn from this study: (1) AGR produces high quality drinking water at all studied sites; the organic carbon content reached the Finnish national recommended value of 2 mg/l TOC. In general, higher quality water is obtained during AGR compared to chemically treated water. (2) Large molecular fractions of NOM are removed more efficiently than smaller fractions. No accumulation of fractionated products in water occurs. (3) The relative content of aromatic carbon compounds increases in sprinkling networks and at some points of infiltration indicating preferential removal of non-aromatic compounds. In the subsurface, however, no clear trend occurs, as relative aromatic carbon content in water varies between the sites and different sampling locations. (4) Total cell counts in extracted groundwater are greater than in chemically treated water, in general. (5) A strong positive correlation exists between organic matter removal and cell counts during infiltration indicating a substantial role of biodegradation in the purification process.

(6) Bacterial communities in raw waters and extracted groundwaters are diverse with changes occurring during infiltration. Also, natural groundwater contains a diverse microbial community that is different from the microbial communities in extracted groundwater in the AGR area. This indicates a clear impact of AGR on subsurface microbial ecology. (7) NOM removal efficiency and the microbial community structure in AGR do not follow seasonal trends.

Acknowledgements This study was funded by Maj and Tor Nessling Foundation and Graduate School of Tampere University of Technology, Finland. Ha¨meenlinna, Jyva¨skyla¨ and Tuusula water works involved in the study and Tampere water works laboratory personnel are acknowledged for fruitful cooperation. Also, we thank Hilda Szabo´, MSc, and Laura McLaughlin, M.Sc., for assisting in HPSEC analyses and English editing, respectively. R E F E R E N C E S

Aiken, G., 2002. Organic matter in ground water. In: Aiken, G., Kuniansky, E.L. (Eds.), US Geological Survey Artificial Recharge Workshop Proceedings, April 2–4. Sacramento, California. Alexander, A., 1965. Biodegradation: problems of molecular recalcitrance and microbial fallibility. Adv. Appl. Microbiol. 7, 35–80. Alfreider, A., Kro¨ssbacher, M., Psenner, R., 1997. Groundwater samples do not reflect bacterial densities and activity in subsurface systems. Water Res. 31 (4), 832–840. Anderson, R.T., Lovley, D.R., 1997. Ecology and biogeochemistry of in situ groundwater bioremediation. Adv. Microb. Ecol. 15, 289–350. Bengtsson, G., 1989. Growth and metabolic flexibility in groundwater bacteria. Microb. Ecol. 18, 235–248. Drewes, J.E., Fox, P., 1999. Fate of natural organic matter (NOM) during groundwater recharge using reclaimed water. Water Sci. Technol. 40 (9), 241–248. Drewes, J.E., Jekel, M., 1998. Behavior of DOC and AOX using advanced treated wastewater for groundwater recharge. Water Res. 32 (10), 3125–3133. Finnish Ministry of Social Affairs and Health, 1994. Guidelines for quality requirements and surveillance of drinking water. Directive 74/94. Flemming, H.-C., 1995. Sorption sites in biofilms. Water Sci. Technol. 32 (8), 27–33. Harvey, R.W., Barber II, L.B., 1992. Associations of free-living bacteria and dissolved organic compounds in a plume of contaminated groundwater. J. Contam. Hydrol. 9 (1–2), 91–103. Hazen, T.C., Jimene´z, L., de Victoria, G.L., Fliermans, C.B., 1991. Comparison of bacteria from deep subsurface sediment and adjacent groundwater. Microb. Ecol. 22, 293–304. Helmisaari, H.-S., Illmer, K., Hatva, T., Lindroos, A.-J., Pa¨a¨kko¨nen, J., Reijonen, R. (Eds.), 2003. Artificial groundwater recharge: infiltration technique, subsurface processes and water quality (Tekopohjaveden muodostaminen: imeytystekniikka, maapera¨prosessit ja veden laatu). Final Report of a Research Project Carried Out in the Finnish Forest Research Institute, 220pp. Hendel, B., Marxsen, J., Fiebig, D., Preuß, G., 2001. Extracellular enxyme activities during slow sand filtration in a water recharge plant. Water Res. 35, 2482–2488.

ARTICLE IN PRESS WAT E R R E S E A R C H

4 1 (200 7) 271 5 – 272 5

Hirsch, P., Rades-Rohkohl, E., 1983. Microbial diversity in a groundwater aquifer in northern Germany. Ind. Microbiol. 24, 183–200. Juhna, T., Klavins, M., Eglite, L., 2003. Sorption of humic substances on aquifer material at artificial recharge of groundwater. Chemosphere 51, 861–868. Kalbitz, K.K., Schwesig, D., Schmerwitz, J., Kaiser, K., Haumaier, L., Glaser, B., Ellerbrock, R., Leinweber, P., 2003. Changes in properties of soil-derived dissolved organic matter induced by biodegradation. Soil Biol. Biochem. 35, 1129–1142. Karanfil, T., Erdogan, I., Schlautman, M.A., 2005. The role of filtration in DOC, UV 254 and SUVA 254 determinations. American Water Works Association Research Foundation, Denver, Colorado, 370pp. Ka¨stner, M., Hofrichter, M., 2001. Biodegradation of humic substances. In: Steinbu¨chel, A. (Ed.), Biopolymers. WileyVCH, Germany. Kilb, B., Kuhlmann, B., Eschweiler, B., Preuß, G., Ziemann, E., Scho¨ttler, U., 1998. Community structures of different groundwater habitats investigated using methods of molecular biology (Darstellung der mikrobiellen besiedlungsstruktur verschiedener grundwasserhabitate durch anwendung molekularbiologischer methoden). Acta hydrochim. hydrobiol. 26 (6), 349–354. Kivima¨ki, A.-L. (Ed.), 2001. Transport and degradation of organic compounds. In: Final Report of EC Project ENV4-CT95-0071 Artificial recharge of groundwater, pp. 109–114. Kortelainen, N.M., Karhu, J.A., 2006. Tracing the decomposition of dissolved organic carbon in artificial groundwater recharge using carbon isotope ratios. Appl. Geochem. 21, 547–562. La˚ngmark, J., Storey, M.V., Ashbolt, N.J., Stenstro¨m, T.A., 2004. Artificial groundwater treatment: biofilm activity and organic carbon removal performance. Water Res. 38, 740–748. Lehtola, M.J., Miettinen, I.T., Vartiainen, T., Martikainen, P.J., 2002. Changes in content of microbially available phosphorus, assimilable organic carbon and microbial growth potential during drinking water treatment processes. Water Res. 36, 3681–3690. Lindroos, A.-J., Kitunen, V., Derome, J., Helmisaari, H.-S., 2002. Changes in dissolved organic carbon during artificial recharge of groundwater in a forested esker in Southern Finland. Water Res. 36, 4951–4958. Madsen, E.L., Ghiorse, W.C., 1993. Groundwater microbiology: subsurface ecosystem processes. In: Ford, T.E. (Ed.), Aquatic Microbiology: An Ecological Approach. Blackwell, Boston, pp. 167–213.

2725

Marschner, B., Kalbitz, K., 2003. Controls of bioavailability and biodegradability of dissolved organic matter in soils. Geoderma 113, 211–235. Miettinen, I.T., Vartiainen, T., Martikainen, P.J., 1996. Bacterial enzyme activities in ground water during bank filtration of lake water. Water Res. 30 (10), 2495–2501. O’Loughlin, E., Chin, Y.-P., 2001. Effect of detector wavelength on the determination of the molecular weight of humic substances by high-pressure size exclusion chromatography. Water Res. 35, 333–338. Pe´rez-Paricio, A., Carrera, J., 2001. Clogging and heterogeneity. In: Final Report of EC Project ENV4-CT95-0071 Artificial recharge of groundwater, pp. 51–58. Peuravuori, J., Pihlaja, K., 1997. Molecular size distribution and spectroscopic properties of aquatic humic substances. Anal. Chim. Acta 337, 133–149. Pre´vost, M., Rompre´, A., Coallier, J., Servais, P., Laurent, P., Cle´ment, B., Lafrance, P., 1998. Suspended bacterial biomass and activity in full-scale drinking water distribution systems: impact of water treatment. Water Res. 32, 1393–1406. Rauch, T., Drewes, J.E., 2005. Quantifying biological organic carbon removal in groundwater recharge systems. J. Environ. Eng. 131, 909–923. Rauch-Williams, T., Drewes, J.E., 2006. Using soil biomass as an indicator for the biological removal of effluent-derived organic carbon during soil infiltration. Water Res. 40, 961–968. Thurman, E.M., 1985. Organic Geochemistry of Natural Waters. Martinus Nijhoff/Dr. W. Junk Publishers, Dordrecht, The Netherlands, 497pp. Tiirola, M.A., Wang, H., Paulin, L., Kulomaa, M.S., 2002. Evidence for natural horizontal transfer of the pcpB gene in the evolution of polychlorophenol-degrading Sphingomonads. Appl. Environ. Microbiol. 68 (9), 4495–4501. van der Kooij, D., Veenendaal, H.R., Baars-Lorist, C., van der Klift, D.W., Drost, Y.C., 1995. Biofilm formation on surfaces of glass and teflon exposed to treated water. Water Res. 29 (7), 1655–1662. Wang, G.-S., Hsieh, S.-T., 2001. Monitoring natural organic matter in water with scanning spectrophotometer. Environ. Int. 26, 205–212. Wuertz, S., Okabe, S., Hausner, M., 2004. Microbial communities and their interactions in biofilm systems: an overview. Water Sci. Technol. 49 (11–12), 327–336. Yavuz Corapcioglu, M., Haridas, A., 1984. Transport and fate of microorganisms in porous media: a theoretical investigation. J. Hydrol. 72 (1–2), 149–169.