Transport of Escherichia coli strains isolated from natural spring water

Transport of Escherichia coli strains isolated from natural spring water

Journal of Contaminant Hydrology 140-141 (2012) 12–20 Contents lists available at SciVerse ScienceDirect Journal of Contaminant Hydrology journal ho...

729KB Sizes 1 Downloads 29 Views

Journal of Contaminant Hydrology 140-141 (2012) 12–20

Contents lists available at SciVerse ScienceDirect

Journal of Contaminant Hydrology journal homepage: www.elsevier.com/locate/jconhyd

Transport of Escherichia coli strains isolated from natural spring water G. Lutterodt a,⁎, J.W.A. Foppen b, S. Uhlenbrook b, c a b c

Department of Civil Engineering, Central University College, P.O. Box DS 2310 Accra,Ghana UNESCO-IHE Institute for Water Education, P.O. Box 3015, 2601 DA Delft, The Netherlands Delft University of Technology, Department of Water Resources P.O. Box 5408, 2600 GA Delft, The Netherlands

a r t i c l e

i n f o

Article history: Received 29 November 2011 Received in revised form 8 July 2012 Accepted 23 August 2012 Available online 30 August 2012 Keywords: Springs Kampala Escherichia coli Attachment efficiency

a b s t r a c t We present a new methodology to scale up bacteria transport experiments carried out in the laboratory to practical field situations. The key component of the methodology is to characterize bacteria transport not by a constant sticking efficiency, but by a range of sticking efficiency values determined from laboratory column experiments. In this study, initially, we harvested six Escherichia coli strains from springs in Kampala, the capital of Uganda, and then we carried out a number of experiments with 1.5 m high columns of quartz sand with various sampling ports in order to determine the fraction of bacteria as a function of sticking efficiency. Furthermore, we developed a simple mathematical formulation, based on the steady-state analytical solution for the transport of mass in the subsurface, to arrive at bacteria concentrations as a function of transport distance. The results of the quartz sand column experiments indicated that the fractional bacteria mass and sticking efficiency of most of the strains we harvested could be adequately described by a power law. When applying the power distributions to the field situation in Kampala, we found that the transport distance required to reduce bacteria concentrations with five log units ranged from 1.5 to 23 m, and this was up to three times more than when using a constant sticking efficiency. The methodology we describe is simple, can be carried out in a spreadsheet, and in addition to parameters describing transport, like pore water flow velocity and dispersion, only two constants are required, which define the relation between sticking efficiency and percentage of bacteria mass. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Globally, groundwater systems provide 25–40% of the world's drinking water (Morris et al., 2003). The importance of the resource is often attributed to the assumption that it is free of pathogenic microorganisms (e.g., Bhattacharjee et al., 2002). However, many water borne disease outbreaks are caused by the consumption of groundwater contaminated by pathogenic microorganisms (Bhattacharjee et al., 2002; Close et al., 2006; Macler and Merkle, 2000; Powell et al., 2003). Traditionally, strategies employed to protect groundwater sources from contamination rely upon effective natural attenuation of sewage-derived microorganisms by soils (and rocks) over set back distances (Taylor et al., 2004). The prediction of transport ⁎ Corresponding author. E-mail address: [email protected] (G. Lutterodt). 0169-7722/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jconhyd.2012.08.011

distances of microorganisms in aquifers has usually been determined with the classical colloid filtration theory (CFT; Tufenkji and Elimelech, 2004a,b; Yao et al., 1971). The theory is based on the assumption that colloid retention follows an invariable rate deposition on collector surfaces, while fluidphase colloid concentrations reduce log-linearly with increasing distance of transport. However, recent research results indicate that the sticking efficiency of a biocolloid population varies due to variable surface properties of individual members of the population, resulting in differences in affinity for collector surfaces (Albinger et al., 1994; Baygents et al., 1998; Foppen et al., 2007a; Li et al., 2004; Lutterodt et al., 2009a; Simoni et al., 1998; Tong and Johnson, 2007; Tufenkji and Elimelech, 2005). Like other researchers (Redman et al., 2001a,b; Tufenkji et al., 2003), we demonstrated in our previous works (Lutterodt et al., 2009b, 2011) that a power-law best describes the distribution of bacteria mass fraction retained in the saturated porous medium

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

2. Materials and methods 2.1. The study area Kampala is underlain by a variety of meta-sedimentary rocks and weathering has produced a pronounced topography. The town has a shallow aquifer in the weathered regolith and site investigations showed the presence of preferential flow paths, although it is not known how far these extend (Howard et al., 2003). Previous studies in Uganda indicated that the top of the regolith is composed of fine material, with increasingly coarser (sandy clay) material found at depth. Productive aquifers are commonly associated with these layers (Howard et al., 2003; Kulabako et al., 2007). Recharge tends to occur during two distinct wet seasons, although Kampala experiences rainfall throughout the year due to its proximity to Lake Victoria (Nyenje et al., 2010). Kampala has many springs, which are typically protected with a concrete slab and a small protected area upstream of the spring (Fig. 1). In Kampala, six E. coli strains were isolated from springs. They were located in Formal Residential areas (E. coli strains named FR02, FR05, and FR08), an urban FArm (FA03) and informal residential areas or SLums (SL03, SL20). 2.2. E. coli sample collection Sampling was undertaken in July 2010. Thereto, 250 mL spring water samples were collected in sterile polypropylene bottles by means of a syringe, and 100 mL of each sample was passed through a 0.45 μm cellulose acetate filter, which was then placed on a plate of Chromucult agar (DifcoTM LB Broth, Miller), and transported to the microbiological laboratory of Makerere University and incubated at 37 °C for 24 h. The purple color of E. coli colonies on the agar plates

0.8

0.6

C/C0

and their corresponding so called segment sticking efficiencies (Lutterodt et al., 2009b, 2011) when transported through columns of saturated quartz sand. Others found a log-normal distribution (Tong and Johnson, 2007; Tufenkji et al., 2003) or a dual distribution (Foppen et al., 2007a,b; Tufenkji and Elimelech, 2004b, 2005). Escherichia coli (E. coli), a gram-negative, facultative nonspore forming, rod shaped bacterium is commonly used as an indicator of fecal contamination of drinking water supplies, because E. coli is a consistent, predominantly facultative inhabitant of the intestinal tract of warm blooded animals. In addition, E. coli is easy to detect and quantify. Furthermore, the net negative surface charge and low inactivation rates of E. coli enable long travel distances in the subsurface (Foppen and Schijven, 2006). Due to the importance of E. coli, considerable attention has been given to understand their transport and fate in saturated porous media (e.g. Bolster et al., 2009; Foppen and Schijven, 2006; Schinner et al., 2010). In this study, we aimed to determine the sticking efficiency distribution of a number of E. coli strains harvested from natural springs in Kampala, the capital of Uganda. Furthermore, we developed a mathematical tool in order to apply the resulting sticking efficiency distributions to estimate realistic transport distances required to reduce bacterial cell concentrations to sufficiently low concentrations, and to reduce the risk of contamination for condition's characteristic for the shallow subsurface of Kampala.

13

0.4

0.2

0.0

0

50

100

150

Time (minutes) Fig 1. Breakthrough curves of Escherichia coli strain SL03 at various sampling ports (see legend) in a 1.5 m saturated quartz sand column.

allowed us to detect E. coli among different bacteria species growing on the agar plates. Later, E. coli cells were confirmed by serotyping at the Dutch National Institute of Public Health and Environmental Hygiene (RIVM), Bilthoven, the Netherlands, where serotyping was performed according to standard procedures (Ewing, 1986; Guinée et al., 1972). A sterile toothpick was used to pick a single colony of E. coli from the agar plate, which was inoculated into 5 mL of nutrient broth followed by incubation at 37 °C for 24 h. Then, 1 to 2 mL of the freshly grown E. coli cells were inoculated into sterile vials (Microbank™-Dry, PRO-LAB DIAGNOSTICS, Toronto, Canada) containing porous beads saturated with cryopreservative (cryovials), which serve as carriers to support the microorganisms. The vials were stored at −70°C and then transported to the UNESCO-IHE laboratory, Delft, the Netherlands.

2.3. Bacteria growth and size measurements To conduct the experiments, sterile forceps were used to remove a bead from the vial, which was put into 25 mL of nutrient broth. The broth was incubated for 24 h at 37 °C, while shaking at 150 rpm on an orbital shaker. Then, 5 mL of the bacteria solution was further inoculated into four Erlenmeyer flasks containing 250 mL of nutrient broth and again grown for 24 h on an orbital shaker at 150 rpm for 24 h to obtain a cell concentration of ~10 9 cells/mL. The bacteria were washed and centrifuged (4600 rpm) three times in Artificial Ground Water (AGW), which was prepared with 526 mg/L CaCl2;2H2O, 184 mg/L MgSO4;7H2O, buffered with 8.5 mg/L KH2PO4, 21.75 mg/L K2HPO4 and 17.7 mg/L Na2HPO4. The final pHvalue of the suspensions ranged from 6.6 to 6.8, while the electrical conductivity ranged from 980 to 1024 μS/cm. To determine the width and length of the E. coli cells, a light microscope (Olympus BX51) in phase contrast mode, with a camera (Olympus DP2) mounted on top and connected to a computer, was used to take images of the cells. Per E. coli strain 30 different images were imported into an image processing program (DP-Soft 2) and the average cell width and cell length were determined. The equivalent spherical diameter (ESD) was determined as the geometric mean of the average length and width (Rijnaarts et al., 1993).

14

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

2.4. Porous media The porous media comprised of 99.1% pure quartz sand (Kristall-quartz sand, Dorsilit, Germany) with sizes ranging from 180 to 500 μm, while the median of the grain size weight distribution was 356 μm. With this median grain size, straining was excluded as a possible retention mechanism in the column set up: assuming a bacteria equivalent spherical diameter of 1.5 μm, the ratio of colloid and grain diameter was 0.004, which was well below the ratio (0.007) for which straining was observed by Bradford et al. (2007) for carboxyl latex microspheres with a diameter of 1.1 mm suspended in solutions with ionic strengths up to 31 mM (the ionic strength of the solutions we used was 4.7 mmol/L only). Total porosity was determined gravimetrically to be 0.38. Prior to the experiments, the sand was rinsed sequentially with acetone, hexane and concentrated HCl, followed by repeated rinsing with de-mineralized water until the electrical conductivity was close to zero (Li et al., 2004). This was done to remove impurities. 2.5. Column experiments To study the transport of the six E. coli strains isolated from the Kampala springs, column experiments were conducted in artificial groundwater (AGW) prepared as described above. The column consisted of a straight tube of 1.5 m transparent acrylic glass (Perspex) with an inner diameter of 9 cm, and with five sampling ports placed at 20–40 cm intervals along the tube. A stainless steel grid for supporting the sand was placed at the bottom of the tube. The column was gently filled with the clean quartz sand under saturated conditions, while the sides of the column were continuously tapped during filling, to avoid layering or trapping of air. The column was connected both at the funnel shaped effluent end and influent end with two Masterflex pumps (Console Drive Barnant Company Barrington Illinois, USA) via silicone tubes, and the pumps were adjusted to a mean fluid velocity of 1.16×10−4 m/s. Prior to a column experiment, the column was flushed for 18 h with AGW to achieve at stable fluid chemistry inside the column. Bacteria influent suspensions were prepared by washing and centrifuging at 4600 rpm for 10 min three times in AGW. Bacteria cell concentrations in the influent suspension were approximately 109 cells/mL. Experiments were conducted by applying a pulse of 0.3 PV (approximately 1.1 L) of bacteria influent suspension to the column, followed by bacteria free AGW. Samples were taken at five distances from the column inlet except for the experiment with SL20 where samples were taken at four sampling distances. Samples were diluted three times in AGW and, to determine cell concentrations, the optical density (OD) was measured at 254 nm using a spectrophotometer (Cecil 1021; Cecil Instruments Inc., Cambridge, England). Thereto, prior to the experiments, cell numbers were deduced from a polynomial equation after calibration with plate counts on Chromocult agar, using the serial dilution method. To do this, seven 10-fold dilutions of an undiluted E. coli suspension were prepared (in triplicate). The OD of all dilutions was measured and the 107 fold dilution was plated in triplicate. The relation between the number of cells and OD was best fitted (R2 =0.99) by a second-order polynomial. Bacteria inactivation was assessed in all experiments by plating samples of the influent at 30 minute intervals during the entire experiment. All plates were incubated at 37 °C for 24 h. After

each experiment, to clean the sand in the column, and to prepare for the next experiment, a pulse of 0.5 L 1.9 M HCl followed by a pulse of 0.5 L 1.5 M NaOH was flushed through the column, followed by flushing with AGW water until the electrical conductivity and pH of the effluent were equal to that of the AGW. 2.6. Segment sticking efficiency and bacteria fraction retained To verify the distribution in cell affinity for collector surfaces, sticking efficiencies within column segments were computed for a fraction of cells retained in a column slice as (Lutterodt et al., 2009a; Martin et al., 1996):   2 dc Mi αi ¼ − ln ð1Þ 3 ð1−θÞη0 Li M i−1 where αi is the dimensionless sticking efficiency of a column slice i, dc is the median grain size (m), η0 is the single collector contact efficiency (−), θ is the total porosity of the sand (−), and Li is the length of the column slice i, i.e. the distance (m) between two sampling ports. Mi−1 is the total number of cells entering slice i, obtained from the breakthrough curve determined at the upper sampling port of slice i and Mi is the total number of cells, obtained from the breakthrough curve determined at the lower sampling port of slice i using (Kretzschmar et al., 1997): t

Mi ¼ q∫ C ðt Þdt

ð2Þ

0

where q is the volumetric flow rate (mL/min), C is the cell suspension (# cells/mL) and t is the time (min). Since the column experiments were carried out with quartz sand, which is negatively charged at pH 6.6–6.8 (the pH value of the AGW), and since E. coli bacteria are also negatively charged (e.g. Foppen et al., 2010), the sticking efficiency we determined with the column experiments was the sticking efficiency for unfavorable attachment sites. The Tufenkji Elimelech (TE) correlation equation (Tufenkji and Elimelech, 2004a) was used to compute η0. Thereto, 1055 kg/m3 was assumed for the bacteria density, while the Hamaker constant was estimated to be 6.5×10−21 J (Walker et al., 2004). The number of retained bacteria in a quartz sand column slice as a fraction of the total number of bacteria cells injected in the column, Fi,was calculated as (Lutterodt et al., 2009b, 2011): Fi ¼

Mi −M i−1 M0

ð3Þ

where M0 is the total number of E. coli cells in the influent suspension. 2.7. Determining the size of the protection area When the flow field in the vicinity of the springs is in steady state, and when we consider the transport of E. coli to be onedimensional and in steady state, then the mass balance for E. coli in the fluid phase, when neglecting decay of E. coli attached to the solid phase, reduces to: D

d2 C dC −ka C−ki C ¼ 0 −v dx dx2

ð4Þ

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

where D is the dispersion coefficient [m2/d], C is the E. coli concentration [number of colonies/100 mL], x is the traveled distance [m], v is the pore water flow velocity [m/d], ka is the attachment rate coefficient (s−1), and ki is the inactivation or decay rate coefficient (s−1). With boundary conditions dC ð∞Þ ¼ 0: dx

ð5Þ

Eq. (4) can be solved analytically (Van Genuchten, 1981):    9 8 1 i ÞD < v  v 1 þ 4ðka þk = 2 x v2 : C ¼ C 0 exp : ; 2D

3ð1−θÞ vη0 α 2ac

3.1. Column experiments

ð7Þ

where α is the dimensionless sticking efficiency, and in case of geochemical heterogeneity, α can be expressed as (Elimelech et al., 2000; Foppen and Schijven, 2005; Johnson et al., 1996): α ¼ λα f þ ð1−λÞα u

ð8Þ

where αf and αu are the sticking efficiencies to favorable and unfavorable attachment sites respectively, and λ is a dimensionless heterogeneity parameter describing the fraction of aquifer grains composed of minerals favorable for attachment or grains coated with favorable attachment patches. These mineral surfaces could be sediment grains coated with iron oxides. In those cases, the favorable sticking efficiency, αf, approaches unity (Elimelech et al., 2000; Foppen and Schijven, 2005), which indicates that each bacteria cell striking the surface of such mineral oxide (determined by η0, the single collector contact efficiency, as described above) will stick to it. From the column experiments, we arrived at a relation between the unfavorable sticking efficiency, αu, and the fraction of the bacteria population, F, possessing this unfavorable sticking efficiency: C2

F ¼ C 1 ðα u Þ

 or α u ¼

F C1

1

C2

ð9Þ

where C1 and C2 are the constants. By making Eq. (9) discrete, and solving Eq. (6) for these discrete portions or fractions of a bacteria population possessing one favorable sticking efficiency (which was always unity) and one unfavorable sticking efficiency, and by assuming values for v, D, and ki, we determined the bacteria concentration as a function of transport distance per homogeneous fraction. The result was an array of bacteria concentrations in x and F, from which we determined the total bacteria concentration, C, at a distance, x, as the sum of the fractional CF values. C ðxÞ ¼ C F1 ðxÞ þ C F2 ðxÞ þ C F3 ðxÞ þ …

3. Results

ð6Þ

The attachment rate coefficient is defined as (Harvey and Garabedian, 1991; Tufenkji and Elimelech, 2004a): ka ¼

rather arbitrary. However, we considered five log removal to be a sufficient reduction of the risk of the springs containing pathogenic micro-organisms, and appropriate for environmental conditions, since maximum E. coli concentrations in waste water are usually within the 10 3–105 colonies/100 mL range (Baxter and Clark, 1984; Canter and Knox, 1985; Foppen and Schijven, 2006).

ð10Þ

Eq. (10) was applied to each strain separately. Finally, we implemented Eqs. (6)–(10) in a spreadsheet, and, per bacteria strain, we defined the size of the protection area as the distance x required to achieve a five log reduction in total bacteria concentration C(x). The choice of a five log reduction was

Bacteria inactivation experiments revealed negligible inactivation for all strains for the entire period of the experiments. For the six strains, the maximum peak breakthrough concentration reduced with increasing transport distance, similar to strain SL03 (Fig. 1). The segment sticking efficiency, αi, reduced with increasing transport distance (Fig. 2). Usually, αi-values were highest in the first column segment, and varied between 0.198 for strain FR02 to 0.03 for strain SL20. Lowest αi-values were observed in the most distant column slices, from a minimum of 0.002 for FR02 to 0.1 for strain FR05. In fact, strain FR05 behaved differently: for this strain, the segment sticking efficiency was invariably around 0.1, and as a result, the strain was almost completely removed from the fluid after passing through the 1.5 m quartz sand column. For comparison, we also determined the sticking efficiency values in the conventional way as a constant for the entire bacteria population by using Eq. (1) for the entire column length of 1.47 m (Table 2). From this table, we concluded that the sticking efficiency values calculated in this way ranged from 0.01 to 0.11, and were generally in between the minimum and maximum segment sticking efficiencies. Finally, for most of the strains, the relation between Fi and αi could be well described with a power law distribution function (Table 1) with coefficients of determination (R2) ranging from 0.77 (SL03) to 0.99 (FA03). 3.2. Spring protection area Parameter values used to calculate the spring protection areas are given in Table 3. The hydraulic gradient was computed from topographic data, based on the assumption

1.E+00 1.E-01

Fi (-)

C ð0Þ ¼ C 0 and

15

1.E-02 1.E-03 1.E-04 1.E-05 1.E-03

1.E-02

1.E-01

1.E+00

αi (-) Fig. 2. Fraction of Escherichia coli input mass (Fi) retained in a column segment as a function of segment sticking efficiency (αi).

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

Table 1 Power law functions of Escherichia coli mass fraction retained (Fi) in a saturated quartz sand column segment and the segment sticking efficiencies (αi). Strain

Power law

R2(−)

SL03 FR08 FR02 FA03 SL20

Fi = 8.78αi1.22 Fi = 100.53αi1.86 Fi = 1.11αi1.35 Fi = 10.29αi1.47 Fi = 68.46αi1.58

0.76 0.96 0.84 0.91 0.99

that topographic gradient is generally a good indicator for the groundwater gradient (Tóth, 1963). We carried out two sets of calculations (Fig. 3A and B): one set for a geochemical heterogeneity parameter of 1%, indicating that 99% of the collector surface was negatively charged and one set when heterogeneity was absent (100% negatively charged collector surface). In addition, we determined the concentration as a function of transport distance for the case when αu, the unfavorable sticking efficiency, was constant, whereby we chose the lowest value (0.01) from Table 2 (dashed lines in Fig. 3A and B). For the case when the geochemical heterogeneity parameter was 1%, for all strains, transport distances belonging to a five log removal ranged from 1.5 to 2.5 m. Also, when using a constant unfavorable sticking efficiency of 0.01, the transport distance was around 2.5 m. When geochemical heterogeneity was absent, however, maximum transport distances ranged from 8 to 23 m, while the transport distance using a constant unfavorable sticking efficiency was around 7 m. From the figure, we concluded that compared to using a constant sticking efficiency value (in Fig. 3B compare dashed curve with SL20), our approach with distributed sticking efficiencies yielded around three times higher transport distances. However, when comparing Fig. 3A and B, the influence of geochemical heterogeneity on maximum transport distance was more dominant than the influence of a sticking efficiency distribution. Finally, the fraction of bacteria contributing to the bacteria concentration determined at the maximum transport distance usually possessed an unfavorable sticking efficiency in the order of 10−4–10−3 (Fig. 4); bacteria with higher unfavorable sticking efficiencies were mostly removed, while the mass of bacteria with very low sticking efficiencies (10−4–10−6) was too low to contribute substantially to the bacteria concentration measured at the maximum transport distance.

Table 3 Overview of parameter values used to determine the spring zone protection areas in Kampala. Parameter

Value

Dispersivity

10% of max. travel Appelo and Postma, distance 2005 0.08 Topographic maps

Hydraulic gradient Permeability Porosity Pore water flow velocity Mean sediment grain size Heterogeneity parameter Decay rate coefficient

Source

5 m/d 0.23 1.79 m/d

Kulabako et al. (2007) Kulabako et al. (2007) –

250 μm

Assumption

1%

Assumption

0.25 d

−1

Foppen and Schijven (2006)

4. Discussion The most important conclusion from our work was that, to our knowledge, we were the first to present a simple and straight forward methodology to scale up bacteria transport experiments carried out in the laboratory to a practical field situation. The essence was that we included a range of sticking efficiencies within one and the same strain dependent bacteria population. Key component of our methodology was the use of well defined column experiments with various sampling ports in order to determine the fraction of bacteria as a function of sticking efficiency, and since our column was made of quartz 1.E+00

A 1.E-01

C/CT (-)

16

1.E-02 1.E-03 1.E-04 1.E-05 0

2

4

Distance (m) 1.E+00

B Table 2 Conventional sticking efficiency values calculated with Eq. (1) for the entire column length. Strain

Mass input (#cells)

Mass output (#cells)

α1.47

SL03 FR08 FR02 FR05 FA03 SL20*

1.18E + 12 1.30E + 12 1.18E + 12 1.04E + 12 2.47E + 12 1.18E + 12

4.03E + 11 5.50E + 11 1.90E + 10 8.00E + 10 2.83E + 11 7.70E + 11

0.03 0.02 0.11 0.09 0.05 0.01



C/CT (-)

1.E-01 1.E-02 1.E-03 1.E-04

α1.07 (Sticking efficiency was calculated at a distance of 1.07 m for this particular strain).

1.E-05 0

10

20

30

Distance (m) Fig. 3. Relative strain dependent E. coli concentration as a function of transport distance in the presence of geochemical heterogeneity (A), and absence of geochemical heterogeneity (B).

CF (at x =x)

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

17

1.E+00

4.1. Spring protection in the Kampala area

1.E-01

We have data from Kampala on the geo-available (as defined by Pettersen and Hertwich, 2008) iron fraction, indicating that the average weight percentage of iron was 0.5–1.0%. Although we cannot determine from this the percentage of collector surface area available for favorable attachment, at least we think the shallow aquifer in Kampala is characterized by geochemical heterogeneity. Therefore, when a plume of E. coli would enter the pristine aquifer, it is likely that most of the bacteria would be removed according to Fig. 3A, and spring protection areas of a few meter would be more than adequate. However, the aquifer is not pristine any more, and has been in contact with contaminated water for at least a number of years. Based on earlier work (e.g. Foppen and Schijven, 2005; Foppen et al., 2008), we think a part or most of these positively charged attachment sites will have been occupied by E. coli, remainders of E. coli, or by humic acids, neutrals, and bases, polysaccharides, polyphenols, proteins, lipids, and heterogeneous organic molecules (Fujita et al., 1996; Imai et al., 2002; Ma et al., 2001), usually present in waste water, and likely being present at least to a certain extent in the faecally contaminated groundwater of the shallow aquifer in Kampala. Therefore, in time and upon prolonged infiltration of waste water, favorable attachment will become less and less important, and bacteria transport will be enhanced, ultimately leading to the situation similar to the absence of geochemical heterogeneity (Fig. 3B). We can only speculate whether springs in Kampala are faced with this situation. Although we have not focused on the decay rate coefficient, this parameter, together and in combination with the pore water flow velocity also influences the transport distance. As an example we included Fig. 5, whereby we reduced the decay rate coefficient with 50% compared to Fig. 3A. The results indicated that the transport distance generally increased with a factor 2. Finally, we would like to address the differences between growth conditions of the E. coli strains we used in laboratory and in the Kampala situation. A number of workers have demonstrated the effect of different growth conditions on the attachment of E. coli during transport in columns. For instance, Lutterodt et al. (2011) found that their E. coli sticking efficiency values were lower than that of similar strains they had used before (Lutterodt et al., 2009a), which they attributed to differences in growth conditions. Apparently, growing E. coli in nutrient broth at 37 °C yielded gave rise to lower sticking efficiencies than when growing E. coli strains in a low nutrient cow manure extract at 21 °C. In addition, Yang et al. (2006, 2008), when mimicking (rich) intestinal and (stressed) external environmental conditions during growing of E. coli strains, reported significantly higher expression of cell properties (e.g. Ag43 expression, hydrophobicity and biofilm formation) and higher retention on biobarriers (Yang et al., 2008) for strains grown under external environmental conditions than when grown under intestinal conditions. Although this aspect of the effect of growth conditions on E. coli certainly deserves more attention, these results make clear that the rich growth conditions we used may have resulted in a low expression of those cell surface properties that influence cell attachment to abiotic surfaces, causing relatively low bacteria retention by the quartz grains. For comparison, we inserted the power law distributions obtained from the five E. coli strains grown in a cow manure extract by Lutterodt et al. (2009a) into

1.E-02 1.E-03 1.E-04 1.E-05 1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

unfavorable sticking eff. Fig. 4. Fractional bacteria concentration determined at the maximum transport distance as a function of unfavorable sticking efficiency.

grains, we determined the bacteria fraction as a function of the unfavorable sticking efficiency. In our opinion it does not really matter how long the column is, as long as there is a possibility to sample at various transport distances in order to determine the constants C1 and C2 in the power law equation (Eq. (9)). The observation that the six E. coli strains revealed variation in their interaction with quartz sand, as witnessed by the varying segment sticking efficiencies, αi, is consistent with observations made for E. coli strains isolated from different sources (Bolster et al., 2006, 2009; Lutterodt et al., 2009b, 2011; Schinner et al., 2010). Such reductions in αi with increasing transport distance have been explained in the literature (Baygents et al., 1998; Lutterodt et al., 2009a, 2009b, 2011; Simoni et al., 1998; Tufenkji and Elimelech, 2004a, 2005) by heterogeneity in cell surface characteristics. Many workers have attributed higher sticking efficiencies obtained at transport distances near the influent end of a column to preferential removal of cells with characteristics that promote their attachment to collector surfaces (Bolster et al., 1999, 2000; Foppen et al., 2007a, 2007b; Li et al., 2004). The low aspect ratio between the average collector grain size and average size of E. coli used enabled us to rule out the possibility of straining (see the Materials and methods section). The observed αi variations of 1.1 log-unit within the E. coli strains were consistent with our previous work, despite the fact that growth conditions of the E. coli strains applied in the two studies differed. We can conclude from this that for E. coli strains isolated from different sources, though the magnitude of cell attachment differed; the transport behavior was similar with respect to intra-strain attachment variations. The fact that the relation between Fi and αi for most of the strains could be described by a power law is consistent with our earlier results (Lutterodt et al., 2009b, 2011); and with the results of others (e.g. Brown and Abramson, 2006; Redman et al., 2001a, 2001b; Tufenkji et al., 2003). However, not all sticking efficiency distributions could be described by a power law. Strain FR05 showed a very different behavior, whereby the unfavorable sticking efficiency for all segments of the column was invariably around 0.1. As we have indicated before (Foppen et al., 2010), it was likely that the outer surface membrane of strain FR05 could be held responsible for this behavior, but we do not know exactly which component of the membrane had such affinity for the abiotic quartz grain surface.

18

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

1.E+00

1.E+00

1.E-01

1.E-01

C/CT (-)

C/CT (-)

A 1.E-02 1.E-03

1.E-02 1.E-03 1.E-04

1.E-04

1.E-05

1.E-05 0

10

20

0

30

2

Distance (m) Fig. 5. Relative strain dependent E. coli concentration as a function of transport distance in the absence of geochemical heterogeneity and a decay rate coefficient of 0.125 d−1.

5. Conclusions • The most important conclusion from our work was that, to our knowledge, we were the first to present a practical methodology to scale up bacteria transport experiments carried out in the laboratory to a practical field situation, whereby we included a range of sticking efficiency variations within one and the same bacteria population (e.g. a strain). • The segment sticking efficiency, αi, of the six E. coli strains harvested from various springs in Kampala, Uganda, was not a constant, but reduced with increasing transport distance. • The relation between the fraction of bacteria, Fi, and segment sticking efficiency, αi was adequately described by power law distribution functions. • Transport distances needed for five log removal of bacterial cells ranged from 3 to 22 m, dependent on the presence of geochemical heterogeneity, the distribution of the unfavorable

1.E+00

B 1.E-01

C/CT (-)

Eqs. (6)–(10) for similar conditions as given in Table 2 (see Fig. 6A and B). When comparing Figs. 3 and 6, we concluded that in Fig. 6 most of the strains were completely removed within a few meter, except for UCFL-94, which was, for unfavorable attachment conditions, five log reduced within a transport distance of 23 m. This was in the same range as the results we obtained in our present work. What is the current practice in Kampala? From field observations and observations made by other researchers (Haruna et al., 2005; Kulabako et al., 2007; Nsubuga et al., 2004), many of the protected springs in Kampala lack proper and adequate protection from potential anthropogenic pollutant sources. In some cases, the masonry protecting the spring is faulty, while in other cases the area perimeter fences are absent or broken giving rise to animal access within a few meter from the springs. In addition, we observed on many occasions pit latrines within a distance of 30 m from the spring, and as an extreme example, Nsubuga et al. (2004) reported a separation distance between spring and pit latrine of 0.5 m! These observations coupled with the results we obtained in estimating distances required for a significant reduction in bacterial concentration in spring water samples indicated that, in our opinion, spring protection in the Kampala area was inadequate, and contamination of the springs by fecal matter might have occurred within the neighborhood of the springs, likely at distances less than 25 m from the springs.

4

Distance (m)

1.E-02 1.E-03 1.E-04 1.E-05 0

10

20

30

Distance (m) Fig. 6. Relative strain dependent E. coli concentration as a function of transport distance in the presence of geochemical heterogeneity (A), and absence of geochemical heterogeneity (B). The E. coli strains and the power law distributions were taken from Lutterodt et al. (2009b), in which the E. coli strains were grown in an extract of filter-sterilized cow manure in order to mimic environmental conditions.

sticking efficiency, pore water flow velocity, and the decay rate coefficient. Acknowledgments Our sincere gratitude goes to Dr. Robinah Kulabako of the Public Health and Environmental Engineering Department of the faculty of Technology, University of Makerere, Uganda. Appreciation also goes to Peter Kiyaga who was so helpful in finding the springs in the Kampala area. We would also like to thank the entire laboratory staff of UNESCO-IHE for their immense support during the experiments. This research was funded by the Netherlands Ministry of Development Cooperation (DGIS) through the UNESCO-IHE Partnership Research Fund. It was carried out jointly by UNESCO-IHE, Makerere University, and the Kampala City Council in the framework of the Research Project ‘Addressing the Sanitation Crisis in Unsewered Slum Areas of African Mega-cities’ (SCUSA). It has not been subjected to peer and/or policy review by DGIS, and, therefore, does not necessarily reflect the view of DGIS. References Albinger, O., Biesemeyer, B.K., Arnold, R.G., Logan, B., 1994. Effect of bacterial heterogeneity on adhesion to uniform collectors by monoclonal populations. FEMS Microbial Letters 124, 321–326.

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20 Appelo, C.A.J., Postma, D., 2005. Geochemistry, Groundwater and Pollution. A.A. Balkema Publishers, Leiden, The Netherlands. Baxter, K.M., Clark, L., 1984. Effluent recharge. The effects of effluent recharge on groundwater quality. Technical Report, 199. Water Research Centre, United Kingdom. Baygents, J.C., Glynn, J.R., Albinger, O., Biesemeyer, B.K., Ogden, K.L., Arnold, R.G., 1998. Variation of surface charge density in monoclonal populations: implications for transport through porous media. Environmental Science and Technology 32, 1596–1603. Bhattacharjee, S., Ryan, J.N., Elimelech, M., 2002. Virus transport in physically and geochemically heterogeneous subsurface porous media. Journal of Contaminant Hydrology 57, 161–187. Bolster, C.H., Mills, A.L., Hornberger, G.M., Herman, J., 1999. Spatial distribution of deposited bacteria following miscible displacement experiments in intact sediment core. Water Resources Research 35, 1797–1807. Bolster, C.H., Mills, A.L., Hornberger, G.M., Herman, J., 2000. Effect of intrapopulation variability on the long distance transport of bacteria. Ground Water 38, 370–375. Bolster, C.H., Walker, S.L., Cook, K.L., 2006. Comparison of Escherichia coli and Campylobacter jejuni transport in saturated porous media. Journal of Environmental Quality 35, 1018–1025. Bolster, C.H., Haznedaroglu, B., Walker, S.L., 2009. Diversity in cell properties and transport behaviour among 12 different environmental Escherichia coli isolates. Journal of Environmental Quality 38, 465–472. Bradford, S., Torkzaban, S., Walker, S.L., 2007. Coupling physical and chemical mechanisms of colloid straining in saturated porous media. Water Research 41, 3012–3024. Brown, D.G., Abramson, A., 2006. Collision efficiency distribution of a bacterial suspension flowing through porous media and implications for field transport. Water Research 40, 1591–1598. Canter, L.W., Knox, R.C., 1985. Septic Tank System Effects on Groundwater Quality. Lewis Publishers, Inc., Chelsea, Michigan USA0-87371-012-6. Close, E.M., Pang, L., Flintoft, M.J., Sinton, L.W., 2006. Distance and flow effects on microsphere transport in a large gravel column. Journal of Environmental Quality 35, 1204–1212. Elimelech, M., Nagai, M., Ko, C.-H., Ryan, J., 2000. Relative insignificance of mineral outer surface potential to colloid transport in geochemically heterogeneous porous media. Environmental Science and Technology 34, 2143–2148. Ewing, W.H., 1986. Edwards and Ewing's Identification of Enterobacteriaceae, 4th edition. Elsevier Publishing Company, New York, p. 130. Foppen, J.W., Schijven, 2005. Transport of Escherichia coli in columns of geochemically heterogeneous sediment. Water Research 39, 3082–3088. Foppen, J.W., Schijven, 2006. Evaluation of data from the literature on the transport and survival of Escherichia coli and thermotolerant coliforms in aquifers under saturated conditions. Water Research 40, 401–426. Foppen, J.W.A., van Herwerden, Schijven, J., 2007a. Transport of Escherichia coli in saturated porous media: dual mode deposition and intrapopulation heterogeneity. Water Research 41, 1743–1753. Foppen, J.W.A., van Herwerden, M., Schijven, J.F., 2007b. Measuring and modeling of straining of Escherichia coli in saturated porous media. Journal of Contaminant Hydrology 93, 236–254. Foppen, J.W.A., van Herwerden, M., Kebtie, M., Noman, A., Schijven, J.F., Stuyfzand, P.J., Uhlenbrook, S., 2008. Transport of Escherichia coli and solutes during waste water infiltration in an urban alluvial aquifer. Journal of Contaminant Hydrology 95, 1–16. Foppen, J.W., Lutterodt, G., Roling, W.F.M., Uhlenbrook, S., 2010. Towards understanding inter-strain attachment variations of Escherichia coli during transport in saturated porous quartz sand. Water Research 44, 202–1212. Fujita, Y., Ding, W.-H., Reinhard, M., 1996. Identification of wastewater dissolved organic carbon characteristics in reclaimed wastewater and recharged groundwater. Water Environment Research 68, 867–876. Guinée, P.A.M., Agterberg, C.M., Jansen, W.J., 1972. Escherichia coli O antigen typing by means of a mechanized micro technique. American Society for Microbiology 24, 127–131. Haruna, R., Ejobi, F., Kabagambe, E.K., 2005. The quality of water from protected springs in Katwe and Kisenyi parishes, Kampala city, Uganda. African Health Sciences 5, 14–20. Harvey, R.W., Garabedian, S.P., 1991. Use of colloid filtration theory in modeling movement of bacteria through a contaminated sandy aquifer. Environmental Science and Technology 25, 178–185. Howard, G., Pedley, S., Barrett, M., Nalubega, M., Johal, K., 2003. Risk factors contributing to microbiological contamination of shallow groundwater in Kampala, Uganda. Water Research 37, 3421–3429. Imai, A., Fukushima, T., Matsuhige, K., Kim, Y.-H., Choi, K., 2002. Characterization of dissolved organic matter in effluents from wastewater treatment plants. Water Research 36, 859–870. Johnson, P.R., Sun, N., Elimelech, M., 1996. Colloid transport in geochemically heterogeneous porous media; modeling and measurements. Environmental Science and Technology 30, 3284–3293.

19

Kretzschmar, R., Barmettler, K., Grolimund, D., Yan, Y.-D., Borkovec, M., Sticher, H., 1997. Experimental Determination of Colloid Deposition Rates and Collision Efficiencies in Natural Porous Media. Kulabako, N.R., Nalubega, M., Thunvik, R., 2007. Study of the Impact of Land Use and Hydrogeological Settings on the Shallow Groundwater Quality in a Peri-urban Area of Kampala, Uganda. Li, X., Scheibe, T.D., Johnson, W.P., 2004. Apparent decreases in colloid deposition rate coefficients with distance of transport under unfavourable deposition conditions: a general phenomenon. Environmental Science and Technology 38, 5616–5625. Lutterodt, G., Basnet, M., Foppen, J.W.A., Uhlenbrook, S., 2009a. The effect of surface characteristics on the transport of multiple Escherichia coli isolates in large scale columns of quartz sand. Water Research 43, 595–605. Lutterodt, G., Basnet, M., Foppen, J.W.A., Uhlenbrook, S., 2009b. Determining minimum sticking efficiencies of six environmental Escherichia coli isolates. Journal of Contaminant Hydrology 110, 110–117. Lutterodt, G., Foppen, J.W.A., Maksoud, A., Uhlenbrook, S., 2011. Transport of Escherichia coli in 25 m quartz sand columns. Journal of Contaminant Hydrology 119, 80–88. Ma, H., Allen, H.E., Yin, Y., 2001. Characterization of isolated fractions of dissolved organic matter from natural waters and a wastewater effluent. Water Research 35, 985–996. Macler, B.A., Merkle, J.C., 2000. Current knowledge on groundwater microbial pathogens and their control. Hydrogeology Journal 8, 29–40. Martin, M.J., Logan, B.E., Johnson, W.P., Jewett, D.G., Arnold, R.G., 1996. Scaling bacteria filtration rates in different sized porous media. Journal of Environmental Engineering 122, 407–415. Morris, B.L., Lawrence, A.R.L., Chilton, P.J.C., Adams, B., Calow, R.C., Klinck, B.A., 2003. Groundwater and its susceptibility to degradation. A Global assessment of the problem and options for management. : Early Warning and Assessment Report Series, RS. 03–3. United Nations Environment Programme/DEWA, Nairobi, Kenya. Nsubuga, F.B., Kansiime, F., Okot-Okumu, J., 2004. Pollution of protected springs in relation to high and low density settlements in Kampala— Uganda. Physics and Chemistry of the Earth 29, 1153–1159. Nyenje, P.M., Foppen, J.W., Uhlenbrook, S., Kulabako, R., Muwanga, A., 2010. Eutrophication and nutrient release in urban areas of sub-Saharan Africa — a review. Science of the Total Environment 408, 447–455. Pettersen, J., Hertwich, E.G., 2008. Critical review: life-cycle inventory procedures for long-term release of metals. Environmental Science and Technology 42 (13), 4639–4647. Powell, K.P., Taylor, R.G., Cronin, A.A., Barrett, M.H., Pedley, S., Sellwood, J., Trowsdale, S.A., Lerner, D.N., 2003. Microbial contamination of two urban sandstone aquifers in the UK. Water Research 37, 339–352. Redman, J., Grant, S.B., Olson, T.M., Estes, M.K., 2001a. Pathogen filtration, heterogeneity, and potable reuse of wastewater. Environmental Science and Technology 35, 1798–1805. Redman, J.A., Estes, M.K., Grant, S.B., 2001b. Resolving macroscale and microscale heterogeneity in virus filtration. Colloids and Surfaces A: Physicochemical and Engineering Aspects 191, 57–70. Rijnaarts, H., Norde, W., Bouwer, E.J., Yklema, J.I., Zehnder, A.J.B., 1993. Bacterial adhesion under static and dynamic conditions. Applied and Environmental Microbiology 59, 3255–3265. Schinner, T., Letzner, A., Liedtke, S., Castro, F.D., Eydelnant, I.A., Tufenkji, N., 2010. Transport of selected bacteria pathogens in agricultural soil and quartz sand. Water Research 44, 1182–1192. Simoni, S.F., Harms, H., Bosma, T.N.P., Zehnder, A.J.B., 1998. Population heterogeneity affects transport of bacteria through sand columns at low flow rates. Environmental Science and Technology 32, 2100–2105. Taylor, R., Cronin, A., Pedley, S., Barker, J., Atkinson, T., 2004. The implication of groundwater velocity variations on microbial transport and well head protection-review of field evidence. FEMS Microbiology Ecology 49, 17–26. Tong, M., Johnson, W., 2007. Colloid population heterogeneity drives hyperexponential deviation from classic filtration theory. Environmental Science and Technology 41, 493–499. Tóth, J., 1963. A theoretical analysis of groundwater flow in small drainage basins. Journal of Geophysical Research 68 (16), 4795–4812. Tufenkji, N., Elimelech, M., 2004a. Correlation equation for predicting singlecollector efficiency in physicochemical filtration in saturated porous media. Environmental Science and Technology 38, 529–536. Tufenkji, N., Elimelech, M., 2004b. Deviation from the classical colloid filtration theory in the presence of repulsive DLVO interactions. Langmuir 20, 10818–10828. Tufenkji, N., Elimelech, M., 2005. Breakdown of colloid filtration theory: role of the secondary energy minimum and surface charge heterogeneities. Langmuir 21, 841–852. Tufenkji, N., Redman, J.A., Elimelech, M., 2003. Interpreting deposition patterns of microbial particles in laboratory-scale column experiments Environ. Particulate Science and Technology 37, 616–623.

20

G. Lutterodt et al. / Journal of Contaminant Hydrology 140-141 (2012) 12–20

Van Genuchten, M., 1981. Analytical solutions for chemical transport with simultaneous adsorption, zero-order production and first-order decay. Journal of Hydrology 49, 213–233. Walker, S.L., Redman, J.A., Elimelech, M., 2004. Role of cell surface lipopolysaccharides in Escherichia coli K12 adhesion and transport. Langmuir 20, 7736–7746. Yang, H.-H., Morrow, J.B., Grasso, D., Vinopal, R.T., Smets, B.F., 2006. Intestinal versus external growth conditions change surficial properties in a

collection of environmental Escherichia coli isolates. Environmental Science and Technology 40, 6976–6982. Yang, H.-H., Morrow, J.B., Grasso, D., Vinopal, R.T., Dechesne, A., Smets, B., 2008. Antecedent growth conditions alter retention of environmental Escherichia coli isolates in transiently wetted porous media. Environmental Science and Technology 42, 9310–9316. Yao, K., Habibian, M.T., O'Melia, C.R., 1971. Water and waste filtration: concepts and applications. Environmental Science and Technology 5, 1105–1112.