Environ. Sci. Technol. 2010, 44, 2426–2432
Virus’ (MS2, OX174, and Aichi) Attachment on Sand Measured by Atomic Force Microscopy and Their Transport through Sand Columns RAMESH ATTINTI,† JIE WEI,‡ KALMIA KNIEL,‡ J. THOMAS SIMS,† AND Y A N J I N * ,† Department of Plant and Soil Sciences and Department of Animal and Food Sciences, University of Delaware, Newark, Delaware 19716
Received October 22, 2009. Revised manuscript received January 29, 2010. Accepted February 18, 2010.
Atomic force microscopy (AFM) was used to study the attachment of φX174, MS2, and Aichi viruses on sands of different surface properties: oxide-removed (clean), goethitecoated, and aluminum oxide-coated. Interaction forces between viruses and sand surfaces were measured by contact mode AFM using tips coated with particles of each virus. Column experiments were conducted to quantify the macroscopic transport and retention of the viruses in sand. The average adhesion force measured with AFM was highest between aluminum oxide-coated sand and all three viruses, followed by goethitecoated sand, and was significantly lower on oxide-removed sand. Among the viruses, adhesion on goethite-coated and aluminum oxide-coated sands followed the order of MS2 > Aichi > φX174, and on oxide-removed sand it was φX174 > Aichi > MS2. Column breakthrough results revealed the same retention trend, which was completely consistent with AFM force measurements. Strong electrostatic attraction and, to a lesser extent, hydrophobic interactions are responsible for the much greater removal of all three viruses observed in the oxidecoated sands compared to the oxide-removed sand. Mass recovery data indicate that the removal of φX174, MS2, and Aichi was largely reversible when eluted with 3% beef extract solution at pH 9.5. The Derjaguin-Landau-Verwey-Overbeek (DLVO) and extended DLVO theories provided correct qualitative predictions on the deposition trend observed in the experiments. This study, to the best of our knowledge, was the first to employ AFM to directly measure interaction forces between viruses and solid surfaces; and it was the first to evaluate the retention and transport behavior of Aichi virus, a human pathogen.
Introduction Groundwater contamination by pathogenic viruses is common in many regions of the U.S. as well as worldwide. Drinking water quality is greatly affected by the nature of interactions between disease-causing viruses in groundwater, soil, and subsurface systems (1). Abbaszadegan et al. (2) * Corresponding author phone: 302-831-6962; fax: 302-831-0605; e-mail:
[email protected]. † Department of Plant and Soil Sciences. ‡ Department of Animal and Food Sciences. 2426
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 7, 2010
analyzed 448 groundwater samples from 35 states throughout the U.S. and found that 31.5% samples were infected with one or multiple pathogenic viruses. Viruses are very small microorganisms that cause a wide range of serious and even deadly diseases in humans. A number of studies have reported that viruses, upon contamination of soil, can survive and migrate long distances where they may contaminate drinking water wells (3–5). This can pose a significant health risk to humans and animals. Hence, a sound fundamental understanding of the factors controlling the fate and transport of viruses in natural environment is necessary for protecting public health. The fate and transport of viruses in soils is variable and dependent on complex interactions between viruses and the soil components (6). Sorption and inactivation of viruses in soils are influenced by soil properties such as texture, organic matter content, metal/metal oxides contents, temperature, soil microbial activity, pH, and certain inorganic salts, as well as virus properties such as virus type and virus physical state (7). Attachment of viruses on solid surfaces mainly depends on surface charge and surface area, ionic strength and pH of the solution, and characteristics of the viruses (8, 9). There have been a number of studies that have investigated the retention and transport of viruses through porous media that contained positively charged solid surfaces (8, 10–13). In natural soils and at pH values close to neutral, metal oxides tend to possess a net positive surface charge, which makes them effective sorbents for negatively charged viruses. Chu et al. (14) studied the effect of soil properties on virus transport and reported that the presence of metal oxides in soils was responsible for enhanced virus sorption and inactivation. Zhao et al. (7) also found that a soil that contained Fe and Al oxides had greater efficiency in removing viruses due to their inactivation by or irreversible sorption on the oxides. The retention and transport behavior of viruses in porous media have been extensively studied in columns, however, such studies do not allow detailed investigation on the interactions that occur between viruses and porous materials. Atomic force microscopy (AFM) is a versatile technique that has been employed to examine a wide variety of systems, including characterization of mechanical properties of materials on micro- and nanoscales, measurement of adhesion forces between surfaces, and probing kinetics of bond strength of biomolecules (15). It has also been used to measure interaction forces between bacteria and solid surfaces (16). To understand the factors that control bacteria adhesion to various solid surfaces, AFM studies have been performed to relate bacteria adhesion to various solid surfaces in terms of surface hydrophobicty (17), surface charge (18), and ionic strength, nutrients, and pH (19). However, no results have been published to date on using AFM to examine the attachment behavior of viruses on solid surfaces, presumably due to difficulties associated with viruses’ very small size. Marshall and colleagues (20) were the first to use the classic Derjaguin-Landau-Verwey-Overbeek (DLVO) theory to describe microbial adhesion to solid surfaces. The classic DLVO theory has since been used to explain microbial attachment behavior (1, 21, 22). There have been some attempts to resolve the discrepancies between DLVO predictions and obtained retention data by considering non-DLVO forces due to hydrophobic interactions into the extended DLVO calculations (23–25). Application of DLVO theory to evaluate virus retention in porous media has been very limited (1, 26). 10.1021/es903221p
2010 American Chemical Society
Published on Web 03/05/2010
FIGURE 1. CLSM images of a silicon nitride tip coated with OX174, MS2, and Aichi virus (a); force-distance curves between viruses (OX174, MS2, and Aichi virus) and sand surfaces (b). Aluminum oxide-coated sand (O retraction; × approach), goethite-coated sand (0 retraction; × approach), oxide-removed sand (∆ retraction; 3 approach). In this study, we used AFM to directly measure interaction forces between selected viruses and solid surfaces. AFM force measurements were evaluated against macroscopically quantified virus retention in sand of different surface reactivity by conducting corresponding column experiments. We selected two bacteriophages (φX174 and MS2) and a human enteric virus (Aichi) and prepared oxide-removed, goethite-coated, and aluminum oxide-coated sands. To the best of our knowledge, this was the first study to examine the transport and retention behavior of Aichi virus, a human pathogen that was first recognized in 1989 in Japan as the cause of oyster-associated nonbacterial gastroenteritis in humans (27). By conducting experiments with Aichi virus under the same conditions with two bacteriophages, this study allowed direct comparison of the behavior between Aichi virus and the bacteriophages, which are commonly used as surrogates for human enteric viruses (HEVs). The DLVO and extended DLVO theories were used to calculate virus interaction force profiles, which were then used to evaluate AFM force measurements and measured virus retention in column experiments.
Materials and Methods Porous Material, Synthesis of Metal Oxide-Coated Sands, and Characterization. Accusand (Unimin, Le Sueur, MN) with the following particle size distribution: 9% of 0.1-0.25 mm, 69.8% of 0.25-0.5 mm, and 21.2% of 0.5 -1.0 mm, was used for column transport experiments. The sand was treated with citrate buffer solution to remove impurities and metal oxides using the procedure described by Chu and Jin (28). The treated sand is referred to as pure sand or oxide-removed sand. Syntheses of goethite or aluminum oxide on oxideremoved sand were carried out according to the methods described by Cheng et al. (29) and Kuan et al. (30), respectively, and the details are provided in Supporting Information. A potentiometric titration method was used to measure the “point of zero salt effect” (PZSE) of goethite-coated and aluminum oxide-coated sands (31). The measured PZSE values for goethite-coated and aluminum oxide-coated sands are 7.9 and 9.1, respectively. Scanning electron microscopy (SEM, S4700, Hitachi High Technologies America, Inc.) was
used to examine the morphology of oxide-removed, goethitecoated, and aluminum oxide-coated sands, and energydispersive X-ray analysis (EDXA) was used to confirm the coating of goethite or aluminum oxide on sand (Figure S1 in Supporting Information). Zeta potentials were measured for all sands to be -39.5 ( 0.7 mV, 4.1 ( 0.2 mV, and 11.5 ( 0.6 mV for oxide-removed, goethite-coated, and aluminum oxide-coated sands, respectively. Additionally, contact angles were also measured (16 ( 1.1°, 62 ( 2.1°, and 70 ( 1.6° for oxide-removed, goethite-coated, and aluminum oxide-coated sands, respectively). Detailed procedures for zeta potential and contact angle measurements can be found in Supporting Information. Viruses, Purification, and Assay. Two bacteriophages, MS2 and φX174 and one human enteric virus, Aichi virus, were selected for measurements of AFM forces and virus retention and transport experiments in columns. MS2 is a single-stranded RNA bacteriophage with a diameter of 24-26 nm and an isoelectric point of 3.9 (32). φX174 is a spherical single-stranded DNA with an isoelectric point of 6.6, and its diameter is 23 nm (33). Aichi virus is a single-stranded positive sense RNA virus with a particle size of 30 nm and has a small round structure (34). The procedure for virus purification is reported in Supporting Information. The two bacteriophages were assayed by plaque-forming-unit (pfu) method (35) and Aichi virus was assayed by 50% tissue culture infective dose, or TCID50 method (36). Only the plates with 10-250 pfus per plate MS2 or φX174 were accepted for quantification and the limit of quantification was set to 10 pfu/plate. For Aichi virus, the detection limit was set to 101.2 TCID50/mL. Force Measurements by Atomic Force Microscopy. A bioscope II AFM (Veeco Instruments Inc.) mounted on an Axiovert 200 inverted fluorescent microscope (Carl Zeiss Inc.) was used to measure virus-sand interaction forces using virus-coated silicon nitride cantilevers (Veeco Instruments Inc.). The procedure for coating viruses on an AFM tip is reported in Supporting Information. Images of the AFM tips coated with the three viruses by confocal laser scanning microscopy (CLSM) are shown in Figure 1a. The virus-coated silicon nitride tips were used to quantify the interaction forces between viruses and oxide-removed sand or metal oxidesVOL. 44, NO. 7, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2427
coated sands. As the tip of the cantilever was brought into and out of contact with a surface, approach and retraction force curves were generated. The forces between virus-coated AFM tip and various sands were measured in contact mode in an artificial groundwater (AGW) buffer solution (IS 0.002 M; pH 7.5) at scan rate of 1 Hz and ramp size of 2 µm. Each average value of measured virus-sand interaction force was taken from 4 measurement locations with 250 measurements at each location. Zero force was chosen where the deflection was independent of the piezo position. The cantilever deflection is then converted into force (F) using the Hooke’s law: F ) -kd, where k is the cantilever spring constant (37). The spring constants of virus-coated cantilevers were experimentally calibrated using the thermal tuning method (38) and the values were found to be similar to the manufacturer reported values. Column Experiments. Virus transport experiments were conducted under continuous, saturated flow conditions using polycarbonate columns (10 cm long and 3.8 cm i.d.). Two sets of experiments were performed with the viruses and various sands (see a summary of conducted experiments in Table S1 in Supporting Information). The first set of experiments was conducted with oxide-removed, goethite-coated, and aluminum oxide-coated sands (100%, the percentage is based on the total weight of the loaded sand in the column). The second set was carried out with 25% (by weight) goethitecoated or aluminum oxide-coated sand mixed thoroughly with 75% oxide-removed sand (referred to as “mixed sand”). The same experimental conditions were used for the two sets of experiments and the experimental setup was similar to that illustrated in Jin et al. (39) (details are provided in Supporting Information). Input solution containing bacteriophages (MS2 and φX174) at 2 × 106 pfu/mL, Aichi virus at 1 × 105 TCID50/mL, and bromide tracer at 25 mg/L was used for all experiments. Beef extract solution (3%, pH 9.5) was used to elute any retained viruses in the column to obtain mass recovery and to examine whether the retained viruses were inactivated or irreversibly sorbed. Beef extract is a high ionic strength solution and has been widely used to detach viruses from various solid surfaces (8, 13).
Theoretical Considerations Classic DLVO Theory. The calculated DLVO (40, 41) interaction force profiles at the employed experimental conditions were used to assess the nature of virus interaction with the various sands. The classic DLVO theory sums van der Waals and electrostatic interactions to determine attractive or repulsive force as a function of separation distance (D) between a virus particle and a surface in aqueous solution. The electrostatic double layer interaction force was calculated using the equation of Hogg et al. (42) and the van der Waals interaction force for sphere-plate interactions using the expression by Gregory (43). Extended DLVO Theory. In previous studies it has been reported that hydrophobic interactions could play a significant role in attachment of hydrophobic and amphiphilic colloids and bacteria (25, 44, 45). Therefore, we used the so-called “extended” DLVO theory by including the contribution of hydrophobic interactions in the classic DLVO theory. Currently there are two methods that have been used to describe hydrophobic interactions quantitatively: a theoretical approach developed by van Oss (46) and an empirical approach by Yoon (47). Bergendahl and Grasso (48) compared the two approaches and found that they gave similar results. In the present study, we selected Yoon’s approach to quantify hydrophobic interactions: φhyd(h) ) -
2428
9
K123τc 6h
(1)
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 7, 2010
TABLE 1. Average Adhesion Forces between Viruses and Sands Measured by AFM average adhesion force (nN) virus
aluminum oxide-coated sand
goethite-coated sand
oxide-removed sand
φX174 MS2 Aichi
8.80 ( 1.77 19.05 ( 2.47 15.26 ( 1.31
5.63 ( 0.76 9.08 ( 0.62 8.24 ( 0.24
1.11 ( 0.32 0.49 ( 0.21 0.72 ( 0.26
where K123 is the hydrophobicity constant for the interaction of viruses and sand in water and can be determined as (49) logK123 ) a
(
)
cos θ1 + cos θ2 +b 2
(2)
where θ1 and θ2 are contact angles of virus and sand, respectively, and a and b are system specific constants. We tested different combinations of a and b values (a ) -7 and b ) -18; a ) -5 and b ) -20), which are reported in the literature (25, 49), to quantify the hydrophobic interactions and found that both combinations yielded very similar results. Hence, we used a ) -7 and b ) -18 in this study. The value of 277 K was used corresponding to the temperature at which all the virus transport experiments were carried out (4 °C). The Hamaker constant values for oxide-removed sand-viruswater, goethite-coated sand-virus-water, and aluminum oxide-coated sand-virus-water systems are 3.42 × 10-21 (50), 6.0 × 10-21 (12), and 2.2 × 10-20 (51), respectively.
Results and Discussion Adhesion Force Measurement by AFM. Figure 1b shows a set of sample interaction forces (both approach and retraction) measured between viruses (φX174, MS2, and Aichi) coated on AFM silicon nitride tips and oxide-removed, goethite-coated, and aluminum oxide-coated sands. The mean adhesion (i.e., retraction) force values, each averaged from 4 locations with 250 measurements at each location, along with standard deviations, are summarized in Table 1. These results show that the adhesion forces of the viruses for oxide-removed sand are lower than for goethite-coated sand or aluminum oxide-coated sand. Overall, the adhesion forces on aluminum oxide-coated sand are the highest for all three viruses. The weak interaction between the viruses and oxideremoved sand is not surprising because both the viruses and oxide-removed sand were negatively charged thus creating electrostatic repulsions. On the other hand, the strong adhesion forces measured between φX174, MS2, and Aichi viruses and metal oxide-coated sands are the result of strong attractions between the negatively charged viruses and positively charged coated sands. Surface properties of the sands were greatly modified after coating with goethite or aluminum oxide: the oxide-removed sand at pH 7.5 in AGW solution had a negative zeta potential value (-39.5 ( 0.7 mV) whereas after coating the zeta potential values changed from negative to positive (goethite-coated sand: 4.1 ( 0.2 mV and aluminum oxide-coated sand: 11.52 ( 0.6 mV, see Table S2). The observed strongest adhesion on aluminum oxide-coated sand corresponds to its highest positive zeta potential value as well as the higher PZSE value (9.1). EDXA measurements confirm that the amount of Al on the coated sand was higher than Fe (Figure S1). Stronger hydrophobic interactions may have also contributed to the strong attractions measured between the viruses and metal oxide-coated sands. As shown in Table S2, in addition to change in zeta potential value, coating with iron or aluminum oxides also modified the oxide-removed sand from being hydrophilic with a contact angle of 16 (
FIGURE 2. Viruses breakthrough in (a) oxide-removed, (b) goethite-coated, and (c) aluminum oxide-coated sands. 1.1° in water to more hydrophobic with contact angles of 62 ( 2.1° and 70 ( 1.6°, respectively. Therefore, the stronger attachment of viruses to aluminum oxide-coated sand is likely the result of both stronger electrostatic attraction and hydrophobic interactions. Previous studies have also reported that both contact angle and zeta potential influence bacteria or colloid adhesion to various solid surfaces (17, 52). As shown in Table 1, for oxide-removed sand, the average adhesion force follows the order: φX174 > Aichi virus > MS2; whereas for both oxide-coated sands, the average adhesion force for MS2 are higher compared to the other two viruses following the order: MS2 > Aichi virus > φX174. The calculated standard deviations for virus-sand force curves (n ) 1000, see Table 1) are low indicating excellent reproducibility of the measurements. The differences among the adhesion of φX174, MS2, and Aichi virus on various sands are likely caused by the different surface properties of the viruses, which are discussed in detail in the next section. Retention and Transport in Columns. The breakthrough curves (BTCs) of bromide in oxide-removed, goethite-coated, and aluminum oxide-coated sands (Figure S2) indicate that there was no preferential flow in any of the columns. The similarity among all the curves also indicates that the flow conditions were similar in all columns and that bromide transport was not affected by goethite or aluminum oxide coating. Figure 2 shows the BTCs of MS2, φX174, and Aichi virus through oxide-removed, goethite-coated, and aluminum oxide-coated sands. φX174 was retained slightly higher (C/ C0 ) 89.39%) when compared with MS2 (C/C0 ) 99.01%) and
Aichi virus (C/C0 ) 99.01%) in the oxide-removed sand (Figure 2a). You et al. (10) have also shown greater removal of φX174 than MS2 in a similar oxide-removed sand and suggested that the higher isoelectric point of φX174 (6.6) than MS2 (3.9) hence the lower net negative charge on φX174 led to weaker electrostatic repulsion between φX174 and the sand. Under the experimental conditions employed in this study, all three viruses and the oxide-removed sand were negatively charged, as indicated by their measured zeta potentials, so that it was not surprising that there was no significant retention of the viruses due to electrostatic repulsion. Figure 2b and c clearly indicate that metal oxides-coated sands prevented the transport of all three viruses. This is because surfaces of the coated sands were positively charged in AGW buffer solution at pH 7.5 whereas viruses were negatively charged, which led to strong attraction hence more retention as compared to on oxide-removed sand. The breakthrough concentration (C/C0) of φX174, MS2, and Aichi virus from goethite-coated sand column was 0.03%, 0.004%, and below detection, respectively. The retentions of φX174 and MS2 were slightly higher on aluminum oxide-coated sand (C/C0 values are 0.017% and 0.001% for φX174 and MS2, respectively) than on goethite-coated sand and Aichi virus was again below detection limit. The breakthrough curves of the viruses from the column packed with mixed sand are shown in Figure 3a and b. The breakthrough concentrations from goethite-coated sand were 82.0, 72.7, and 68.7% for φX174, MS2, and Aichi virus, respectively, whereas from aluminum oxide-coated sand they were 41.5, 0.58, and 0.16% for φX174, MS2, and Aichi virus, respectively. Similar to results from experiments with 100% coated sands, removal of φX174 was again lower than MS2 and Aichi virus by both mixed sands. The results also demonstrate that 25% aluminum oxide-coated sand was much more effective in retaining MS2 and Aichi virus than 25% goethite-coated sand. These results are consistent with the AFM force measurements. The surface properties of φX174 likely caused its lower retention on goethite-coated and aluminum oxide-coated sands than MS2 and Aichi virus. Due to its higher isoelectric point, φX174 has fewer negative charges at pH 7.5 than the other two viruses and thus was retained less on the positively charged metal oxide-coated sands. φX174 also has a lower contact angle than MS2 and Aichi virus, making hydrophobic interactions less likely to contribute to the total interactions. The more complete removal of the Aichi virus compared to the other two bacteriophages may suggest a higher reactivity between this HEV with the metal oxides-coated sands and/ or it may be due to its slightly lower input concentration (1 × 105 TCID50/mL) than the bacteriophages’ (2 × 106 pfu/ mL). Mass Recovery. The recoveries of φX174, MS2, and Aichi virus from all experiments are in general high, ranging from ∼100% in oxide-removed sand to ∼60% in oxide-coated sands (a summary of mass recoveries in all experiments is given in Table S1). This observation indicates that retention of the viruses by the various sands was largely reversible. However, it is important to note that the high reversibility was the result of elution with the high pH and high ionic strength BEX. The retention of these viruses was not reversible when eluted with the AGW buffer solution, which is more representative of soil solution and groundwater. Calculated DLVO Profiles. Interaction energy profiles were calculated using both the classic and extended DLVO theories, and the results show that predictions by both theories are similar (Table S3). Thus, only the force profiles calculated by the extend DLVO are shown in Figure 4. The maximum energy barrier is higher for φX174 (1.1 × 10-2 nN) than MS2 (5.4 × 10-3 nN) and Aichi virus (2.2 × 10-3 nN). It has been reported that deposition of bacteria was minimal VOL. 44, NO. 7, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2429
FIGURE 3. Breakthrough curves of viruses in goethite-coated (a) and aluminum oxide-coated (b) sands (25%) mixed with oxide-removed sand (75%).
FIGURE 4. Extended DLVO force profiles between viruses and sands: (a) oxide-removed sand, (b) goethite-coated sand, and (c) aluminum oxide-coated sand. in the primary energy minimum when the energy barrier value is greater than 5 kT (corresponds to ∼1.8 × 10-3 nN) (53). Although the DLVO calculations also indicate the existence of secondary energy minima, their depths in all cases were too shallow (
