Escherichia coli Disinfection by Electrohydraulic Discharges

Jayaram, S.; Castle, G. S. P.; Margaritis, A. Biotechnol. ...... Piero Gallerano , Alfonsina Ramundo-Orlando , Stefania Romeo , Maria Rosaria Scarfì ...
0 downloads 0 Views 94KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 4139-4144

Escherichia coli Disinfection by Electrohydraulic Discharges W.-K. CHING,† A. J. COLUSSI,† H. J. SUN,‡ K. H. NEALSON,‡ AND M . R . H O F F M A N N * ,† W. M. Keck Laboratories, California Institute of Technology, Pasadena, California 91125, and Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109

We study the survival of single-strain Escherichia coli colonies in aqueous media exposed to 5.5 kV, 90 kA electrohydraulic discharges (EHD). The probability of survival (Pn) of a 4 × 107 cfu mL-1 E. coli population after n consecutive EHDs follows a logit distribution: ln(Pn/ 100 - Pn) ) 1.329 - 1.579 ln n with r2 ) 0.993 that corresponds to lethal doses of LD50 ) 2.2 and LD90 ) 10.5 EHDs. Considering that the reactor is thoroughly mixed during each discharge and that LD50 ) 0.9 values are nearly independent of E. coli concentrations in the range of 2 × 103 e E. coli/cfu mL-1 e 3 × 106, we ascribe the nonexponential Pn decay of single-strain E. coli colonies to a shielding phenomenon where inactive cells protect the successively smaller numbers of viable cells in the EHD. The qualitatively similar concentration dependence observed for survival under 254 nm of radiation, in contrast with the lower resistance of denser colonies to 20 kHz power ultrasound and the delayed onset of extracellular β-Dgalactosidase activity in bacterial populations already decimated by EHDs, support the view that UV radiation is the dominant disinfection agent generated by electrohydraulic discharges.

Introduction Many strategies for the management and reduction of microorganisms in water for human consumption have been devised (1). Examples of these strategies include water filtration by screens and osmosis (2), flocculation and aggregation (3), and UV light irradiation (4, 5). By far, the most common treatment technique used is disinfection by chlorine (6). Advanced disinfection techniques make use of ozone (7, 8), hydrogen peroxide (9), ultrasonic irradiation (10-13), γ irradiation (14-16), semiconductor photocatalysis (17-20), flash photolysis (21), biocidal polymers (22, 23), high-intensity laser photolysis (24-27), and high-voltage pulses (28-30) as possible water treatment processes. But as with any emerging technology, cost and treatment volume limit their range of applicability. The continual search for a more general and cost-effective technique for water disinfection is the motivation for this work. Recent discoveries of new and emerging pathogens surviving in water treatment and distribution systems have added urgency to the difficult and heavily regulated process * Corresponding author phone: (626)395-4391; fax: (626)395-2940; e-mail: [email protected]. † W. M. Keck Laboratories. ‡ Jet Propulsion Laboratory. 10.1021/es010643u CCC: $20.00 Published on Web 09/08/2001

 2001 American Chemical Society

of water disinfection. Among these newly identified fecal pathogens are enterohemorrhagic Escherichia coli and Campylobacter jejuni; enteric viruses such as rotavirus, calicivirus, and astrovirus; and parasites Cryptosporidium parvum, Giardia lamblia, and microsporidia that proliferate in facilities where chlorination is primarily used (31). The standard application of disinfectants with a single mode-of-action approach such as chlorine, chloramine, chlorine dioxide, and UV light are sometimes no match for the microbe’s ability to overcome that selective pressure in both planktonic and biofilm environments (32-34). Recently, water treatment facilities have been using a multiple modeof-action approach by combining several disinfectants, both for cost-cutting reasons and for reducing disinfection byproducts (35-37). This combined approach is more difficult to overcome biologically and more appropriate for water disinfection. The potential use of the electrohydraulic discharge process (EHD) falls under this multiple mode of action approach. EHDs generate hot, localized plasmas strongly emitting highintensity UV light but also produce shock waves that break down aggregates and generate hydroxyl radicals during water photodissociation (38). Whereas, microbial resistance to ultraviolet light is commonly observed, no evidence of immunity to combined photonic and mechanical treatment has yet been found. Electrohydraulic discharge reactors have been used effectively in the past for the treatment of environmentally recalcitrant pollutants offering rapid mineralization rates of chemicals such as 2,4,6-trinitrotoluene, 4-chlorophenol, and 3,4-dichloroaniline (38-41). Other applications include simulation of underwater explosions (42), metal forming (43), rock fragmentation (44), and noninvasive removal of gall bladder stones or extracorporeal shock wave lithotripsy (4547). Disinfection by electrohydraulic discharge (EHD) reactors is a novelty motivated by previous work done at Caltech (45-48). In addition, EHD generated UV radiation and shock wave effects on bacteria were studied 30 years ago but in the context of food science and protection (49-51). Electrohydraulic discharges are generated by the rapid discharge of stored electrical charge across submerged electrodes. The formation of an electrical arc across the spark gap produces a localized plasma region that emits UV and VUV radiation and creates pressure and thermal shocks. Measurements of the UV radiation from electrical discharges in water have been reported to be as high as 200 MW peak radiant power (52), and associated tissue damage has been observed in extracorporeal shock wave lithotripsy (46). We demonstrate that, by taking advantage of the intense UV radiation and the mechanical cell damage produced by the discharges, the electrohydraulic discharge process can be used as an alternative method of municipal water disinfection. In this work, we explore the capabilities of EHDs and report the results of the disinfection of water contaminated by E. coli. These experiments identify the dominant mechanism of inactivation and establish the survival doseresponse curves of single-strain E. coli colonies subjected to EHD.

Experimental Methods Disinfection Devices. The Caltech electrohydraulic discharge reactor (Figure 1) has been described in detail in previous publications (38, 39, 41). In a typical experiment, a 135 µF capacitor bank stored energy (ECB ) 7 kJ) at voltage (VCB ) VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4139

FIGURE 1. Schematic diagram of the Caltech electrohydraulic discharge (EHD) reactor and electronic circuit. 10.2 kV) is discharged through a 4-mm electrode gap into 3.0 L of a 0.01 M phosphate-buffered (pH 7.4) saline (PBS) solution within 40 µs. The peak current of these pulses is about 90 kA. The buffered solution has the proper electrical conductivity for the discharge to flow across the immersed electrodes and provides a medium in which suspended E. coli colonies remain in stasis. Upon charging the electrodes, local ohmic heating creates vapor microbubbles within the gap through which the EHDs occur. The energy utilized for vaporization is estimated at 5 kJ/pulse. Therefore, the energy effectively delivered per pulse is about ECB ) 2 kJ. These discharges simultaneously generate a localized hot spot, traveling shock waves, and UV-VUV emissions. Since water absorbs strongly below 185 nm, thermal and VUV effects are necessarily localized within a small volume (a few milliliters) about the electrode gap region. In contrast, mechanical and UV effects may extend further into the reaction volume. We tested mixing conditions in the reactor during and between pulses by following the dispersal of small volumes of colored solutions (1 mL of 0.01 M MnO4K) injected in the gap region. Without pulsing, the colored solute is slowly homogeneized by diffusion/convection throughout the reactor in about 2 h. In contrast, a single pulse suffices to produce a uniformly colored solution. Assuming that in the latter case mixing is driven by pressure waves generated by the discharge, that these waves travel at the speed of sound (1500 m/s), and that they travel back and forth across the reactor 10 times before dissipating, we estimate that mixing takes about 0.6 ms. Since the shortest interval between consecutive discharges is 10 s, the reaction chamber is thoroughly mixed before each discharge. At typical pulse repetition and sampling rates, a 50 consecutive discharge experiment amounts to 2-ms exposure time, but takes approximately 1 h. Figure 2 shows the reactor employed for ultraviolet/ ultrasound (UV/US) disinfection. UV-only experiments were carried out by irradiating E. coli suspensions (100 mL, contained in a bottle fitted with a 1-in. quartz side window) with 254 nm of light emitted by a 15-W Sterilaire SW germicidal lamp. The lamp irradiance (manufacturer specifications) is 4.5 mW/cm2 at 5 in. away from the lamp. Alternatively, suspensions could be sonicated at 20 kHz by means of a 1/2 in. diameter Ti immersible probe (VibraCell Model VCX400). The ultrasonic probe emitted 11.3 W/cm2 while dipping 5 mm into the suspension. Bacterial Cultures and Procedures. Escherichia coli typestrain ATCC 25922 was used throughout. Stock cultures were deposited on LB agar slants in the dark at 5 °C and maintained 4140

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001

FIGURE 2. Schematic diagram of the combined UV and ultrasound reactor system. at -20 °C in LB broth with 60% glycerol added as cryoprotectant. Prior to each experiment, the E. coli stock was innoculated into sterilized LB broth (1% tryptone, 0.5% yeast extract, and 1% salt) and grown in the dark for 12 h at 37 °C while shaken at 225 rpm. Cells from the exponential growth phase were separated by centrifugation, washed once, and resuspended in 0.01 M PBS (24.72 g of Na2HPO4, 3.60 g of NaH2PO4‚H2O, and 170.0 g of NaCl, pH 7.4). Resuspended samples were appropriately diluted, then equilibrated in the corresponding reactors for 5 min, and finally sterilized by EHD, UV, or ultrasound treatment. Samples drawn at regular intervals were held for less than 2 h at 0 °C prior to incubation and counting. Treated samples were diluted in PBS buffer and spread on Petri dishes covered with LB agar (LB broth with 1.5% agar), and the number of colonies (cfu) was counted after incubation in the dark at 37 °C for 16 h. This procedure excludes all the cells that will not reproduce by incubation, regardless of the damage undergone during disinfection. β-D-Galactosidase Assay. Mechanical damage to E. coli membranes could be inferred from the leakage of cellular material during disinfection. Following Miller (53), E. coli cultures were grown (see above) in the presence of 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) to induce the synthesis of β-D-galactosidase (54). Inactivation experiments were performed on 1.0 × 107 modified cfu mL-1 and assayed for leaked β-D-galactosidase within 30 min of sample collection. Every sample was divided into two portions. The

first portion was centrifuged at 8000 rpm for 10 min; the resulting pellet was separated from the supernatant fluid and lysed with CHCl3 and 1% sodium dodecyl sulfate (SDS). β-D-Galactosidase activity was determined in the lysate as well as in the supernatant liquid. The entire second portion was lysed and analyzed for total β-D-galactosidase activity. The fraction of total enzyme activity in the supernatant of the first portion provides a measure of enzyme leakage as function of the number of EHDs. Statistical Analyses. Plotted dose-response data with error bars are shown as the mean ( the largest variance of triplicate agar plate counts. Reproducibility of the disinfection curves is limited by the plasma channel quality from consecutive discharges, with typical acceptable fluctuations in energy of 10%. The statistical analyses of the tailing doseresponse curves (EHD dose vs inactivation percentages) were determined using SigmaPlot (version 5.0) and Microsoft Excel (version 5.0). The rectified disinfection curves were computed by transforming the abscissa (dose) into logarithmic base 10 units and the ordinate (% viability) into logit units using the mapping

logit(yi) ) log10

(

yi 100 - yi

)

(1)

where yi is the percentual (normalized) viability count. This particular graphing method is standard in toxicology experiments to compute values of lethal doses (LD) where tailing components in dose-response curves are common (55). The resulting linearized data sets were analyzed using SigmaPlot by the least-squares fit method of linear regression with associated standard errors for both y-intercepts and slopes. The values of lethal dosages for 50% viability reduction, denoted by LD50, were obtained by setting yi to 0.5 and solving for the dosage n given the y-intercept and the slope m from each regression. A value of yi ) 0.9 was used to compute LD90. The error in the LD50 versus initial cell concentration (C0) plots was computed as the quadratic sum of the error in the viability measurements and the linear regression.

Results and Discussion Disinfection Kinetics and Mechanism. Figure 3a shows the logarithm of the concentration of viable E. coli cells after n EHDs for a colony having an initial concentration of 4.0 × 107 cfu mL-1, at an electrode gap of 0.4 cm, energy ECB ) 7 kJ, at pH 7.4. The bimodal decay of the bacterial population can be realistically fit by a double exponential (the solid line in Figure 3a) or, alternatively, rectified by casting the data into a log it plot (Figure 3b). Let Pn ) 100 × CFUn/CFU0, the percent survival probability of an E. coli colony after n EHDs (56). The expression

(

ln

)

Pn ) 1.329 - 1.579 ln n 100 - Pn

(2)

obtained from a linear regression analysis of the data in Figure 3b actually provides an excellent fit to the data (r2 ) 0.993). From eq 1, one can evaluate the lethal doses required to inactivate 50% and 90% of the initial population: LD50 ) 2.2 and LD90 ) 10.5, respectively. It has been argued that tailing in bacterial survival kinetics is actually associated with experimental artifacts, such as improper mixing, population heterogeneity, biocide quenching (57, 58), etc. Since we verified that the reactor contents are thoroughly mixed by the mechanical perturbation created by EHD pulses, and taking into account that all pulses are in principle equally intense, one should consider the possibility that succeeding discharges might become less effective because the accumulated bacterial debris attenuates their

FIGURE 3. (a) Disinfection of 3 L of a 4.0 × 107 cfu mL-1 E. coli suspension in 0.01 M PBS at pH 7.4 by 50 consecutive electrohydraulic discharges. Discharges are characterized by ECB ) 7 kJ, spark gap ) 4 mm, and pulse rate ) 0.1 Hz. The solid line is a biexponential fit to the data. (b) A logit plot of the data of panel a. The solid line is the result of a linear regression through the data. Pn is the percent survival probability of the colonies after n discharges. lethality. However, UV absorption measurements in the range of 200-400 nm of E. coli suspensions following treatment by 0, 60, and 100 EHDs failed to show any signs of change (Figure 4). Bacterial heterogeneity is not a likely option either since the E. coli suspensions used in these experiments were prepared by dilution of a single pure culture innoculate. That the tailing behavior is not artifactual but intrinsic to this process can be inferred from the fact that the same LD50 value is obtained over a 3 order of magnitude variation of E. coli initial concentrations (Figure 5). The data shown in Figure 5 were obtained by exposing colonies in the range of 103-107 cfu mL-1 to less than four consecutive EHDs, a condition that minimizes turbidity effects from electrode tip erosion. On the other hand, if the decline of disinfection efficiency were the result of disinfection action being limited by the production of a finite amount of lethal agent, E. coli decay would follow zero-order kinetics, at variance with the biexponential dependence observed in Figure 3a. A similar argument rules out localized disinfection, such as that provided by VUV radiation, hydroxyl radical attack, and heat shock about the electrode gap. The tailing in the dose-response curves suggests a shielding phenomenon. The rationale is that each discharge inactivates a finite viable portion of the population and that dead cells are then thoroughly mixed by mechanical perVOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4141

FIGURE 4. UV absorbance measurements using a Shimadzu UV2101PC scanning spectrophotometer with a 10-cm quartz cuvette of 1/100th dilutions of EHD-generated samples with E. coli in 0.01 M PBS after 0 EHDs (solid line), 60 EHDs (dotted line), and 100 EHDs (dashed line).

FIGURE 5. Number of EHD required to inactivate 50% of the initial population (LD50) vs initial E. coli cell concentration (C0) as obtained from initial disinfection rates. The solid line is an empirical threeparameter exponential growth fit to the data. turbation of the EHD. The suspension is returned to the initial fully mixed state but with lower numbers of viable cells. Since all discharges deliver relatively the same UV dose and the total number of bacteria in each experiment is constant, then each successive discharge inactivates smaller portions of viable bacteria due to shielding by inactive cells. Shielding by suspended solids (59, 60) released from tantalum electrode tip erosion have no physical effect in the EHD since these failed to affect the overall UV absorption spectrum, discussed previously in Figure 4. Lethal Agent Characterization. To establish and separate the mechanisms contributing to EHD biocidal action (61), we performed additional disinfection experiments by means of low-intensity UV lamps and also by ultrasonic irradiation. The results of the UV treatment as a function of initial cell concentration are shown in Figure 6, which qualitatively resembles the behavior observed in Figure 5 for EHD disinfection. Dilute colonies are inactivated at constant and relatively rapid rates up to initial cell concentrations exceeding ∼3 × 108 cfu mL-1. We believe that this threshold behavior in EHD and UV experiments is simply a manifestation of the light penetration dependence on cell suspension concentration whose spectrophotometric absorption relation with initial cell concentration closely matches the threshold observations (Figure 6). 4142

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001

FIGURE 6. LD50 values (in min of UV exposure) for E. coli disinfection by 254 nm of low-intensity light vs initial cell concentration (C0) with exponential fit (solid line). Conditions: 100 mL of E. coli suspensions in 0.01 M PBS at pH 7.4. Spectrophotometric absorbance OD600 of E. coli suspensions vs initial cell concentration is shown on the left axis (circles with dotted line).

FIGURE 7. LD50 values (in min of ultrasound exposure) for E. coli disinfection by 20 kHz of ultrasound vs initial cell concentrations. Conditions: 100 mL of E. coli suspensions in 0.01 M PBS at pH 7.4. Ultrasonic irradiation also has deleterious effects on E. coli colonies. However, in contrast with the inverse dependence of efficiency on cell population observed for UV disinfection, it is easier to inactivate the denser colonies (Figure 7). Hydrodynamic shearing of large biological structures with a minimum of redox effects is the expected outcome of cavitation under 20 kHz ultrasound waves. Thus, for example, high molecular weight polymers are degraded by ultrasound down to a limiting chain length (62, 63). Hence, we conclude that bacterial disinfection by ultrasound involves the mechanical disruption of cell aggregates, which is more efficient in the case of dense colonies, a behavior that is not observed in the EHD process (Figure 5). The extent of mechanical damage undergone by cell membranes under EHD was probed by assaying β-Dgalactosidase in treated E. coli samples. Figure 8 shows that after 30 EHDs, about half of the total β-D-galactosidase is released from the cells. The presence of enzyme in the supernatant represent evidence that cell membranes became permeable to the ∼540 kDa moiety. Since extracellular β-Dgalactosidase activity develops after about 10 EHDs, i.e., when only 8% of the initial E. coli population remains viable (Figure 3b), these experiments exclude mechanical damage as a major mechanism of inactivation.

FIGURE 8. β-D-Galactosidase activity of E. coli vs the number of electrohydraulic discharges. t, in the cell lysate. s, in the supernatant fluid. High-Intensity UV Inactivation. Typical rates of laboratory UV lamp inactivation for light doses between 5 and 8 mJ/cm2 achieve at least a 4-log inactivation (4, 5). However, Figure 3a shows that disinfection by 50 consecutive electrohydraulic discharges results in a 99% or 2-log reduction in the viability of E. coli. This comparison cannot be made. The mechanism of high-intensity UV disinfection in the electrohydraulic discharge process is intrinsically different than in low-intensity UV lamp inactivation. A conservative estimate of the UV light intensity generated by the EHD is as high as 3 × 106 W/cm2. Stepwise multiphoton absorption at UV light intensities higher than 106 W/cm2 generated by picosecond and nanosecond laser irradiation is known to produce nonlinear photoprocesses in nucleic acids (24, 64, 65). These photoprocesses can bypass DNA thymidine dimerization (26) in favor of single- and double-strand breaks (ssb, dsb), interstrand cross-linking, and protection by the cytoplasm (66-68) producing drastic differences on inactivation kinetics. Dose-response curves of this type display similar kinetics to disinfection by pulsed radiolysis and ionizing radiation (69, 70). The details of these discussions and calculations are the subject of a subsequent paper (71). Present data on the kinetics of E. coli disinfection by the EHD process are consistent with a logistic dose-response dependence. The evidence presented rules out damage mediated by mechanical or chemical means and points to UV radiation as the dominant lethal agent. We find that bacterial counts become markedly more difficult to reduce in successive treatments, a phenomenon that we ascribe to shielding by dead cells.

Acknowledgments Financial support for this project was provided by the following agencies: DARPA, ONR, and EPRI.

Literature Cited (1) Clesceri, L. S.; Greenberg, A. E.; Trussell, R. R. Standard Methods for the Examination of Water and Wastewater, 17th ed.; Clesceri, L. S., Greenberg, A. E., Trussell, R. R., Eds.; APHA, AWWA, and WPCF: Washington, DC, 1989. (2) Eisenberg, T. N.; Middlebrooks, E. J. Removal of microorganisms by reverse osmosis. Reverse Osmosis Treatment of Drinking Water; Butterworth Publishers: Stoneham, MA, 1986; p 133. (3) Treweek, G. P. Ph.D. Thesis, California Institute of Technology, 1975. (4) Chang, J. C. H.; Ossoff, S. F.; Lobe, D. C.; Dorfman, M. H.; Dumais, C. M.; Qualls, R. G.; Johnson, J. D. Appl. Environ. Microbiol. 1985, 49, 1361. (5) Harris, G. D.; Adams, V. D.; Sorensen, D. L.; Curtis, M. S. Water Res. 1987, 21, 687.

(6) White, G. C. Handbook of Chlorination and Alternative Disinfectants, 4th ed.; Wiley: New York, 1999. (7) Hunt, N.; Marinas, B. Water Res. 1997, 31, 1355. (8) Zhou, H.; Smith, D. J. Environ. Eng. 1992, 120, 841. (9) White, G. C. Ozone, Peroxone, and AOxPs. Handbook of Chlorination and Alternative Disinfectants; Wiley: New York, 1999. (10) Hua, I.; Thompson, J. E. Water Res. 2000, 34, 3888. (11) Vollmer, A. C.; Kwakye, S.; Halpern, M.; Everbach, E. C. Appl. Environ. Microbiol. 1998, 64, 3927. (12) Allison, D. G.; D’Emanuele, A.; Eginton, P.; Williams, A. R. J. Basic Microbiol. 1996, 36, 3. (13) Miller, D. L.; Thomas, R. M.; Frazier, M. E. Ultrasound Med. Biol. 1991, 17. (14) Steenken, S. Chem. Rev. 1989, 89, 503. (15) Teebor, G. W.; Boorstein, R. J.; Cadet, J. Int. J. Radiat. Biol. 1988, 54, 131. (16) von Sonntag, C. The Chemical Basis of Radiation Biology; Taylor & Francis: London, 1987. (17) Matsunaga, T.; Okochi, M.; Takahashi, M.; Nakayama, T.; Wake, H.; Nakamura, N. Water Res. 2000, 34, 3117. (18) Maness, P.-C.; Smolinski, S.; Blake, D. M.; Huang, Z.; Wolfrum, E. J.; Jacoby, W. A. Appl. Environ. Microbiol. 1999, 65, 4094. (19) Sunada, K.; Kikuchi, Y.; Hashimoto, K.; Fujushima, A. Environ. Sci. Technol. 1998, 32, 726. (20) Ireland, J. C.; Klostermann, P.; Rice, E. W.; Clark, R. M. Appl. Environ. Microbiol. 1993, 59, 1668. (21) Oldham, T. C.; Phillips, D. J. Phys. Chem. B 1999, 103, 9333. (22) Sun, G.; Wheatley, W. B.; Worley, S. D. Ind. Eng. Chem. Res. 1994, 33, 168. (23) Sun, G.; Allen, L. C.; Luckie, E. P.; Wheatley, W. B.; Worley, S. D. Ind. Eng. Chem. Res. 1995, 34, 4106. (24) Gorner, H. J. Photochem. Photobiol. B: Biol. 1994, 26, 117. (25) Angelov, D.; Spassky, A.; Berger, M.; Cadet, J. J. Am. Chem. Soc. 1997, 119, 11373. (26) Masnyk, T. W.; Nguyen, H. T.; Minton, K. W. J. Biol. Chem. 1989, 264, 2482. (27) Nikogosyan, D. N.; Oraevsky, A. A.; Zavilgelsky, G. B. Photobiochem. Photobiophys. 1986, 10, 189. (28) van Heesch, E. J. M.; Pemen, A. J. M.; Huijbrechts, P. A. H. J.; van der Laan, P. C. T.; Ptasinski, K. J.; Zanstra, G. J.; de Jong, P. IEEE Trans. Plasma Sci. 2000, 28, 137. (29) Jayaram, S.; Castle, G. S. P.; Margaritis, A. Biotechnol. Bioeng. 1992, 40, 1412. (30) Mizuno, A. IEEE Trans. Ind. Appl. 1988, 24, 387. (31) Szewzyk, U.; Szewzyk, R.; Manz, W.; Schleifer, K.-H. Annu. Rev. Microbiol. 2000, 54, 81. (32) Lin, Y. E.; Vidic, R. D.; Stout, J. E.; Yu, V. L. J. Am. Water Works Assoc. 1998, 90, 112. (33) LeChevallier, M. W.; Cawthon, C. D.; Lee, R. G. Appl. Environ. Microbiol. 1988, 54, 649. (34) Clark, T. F. J. Am. Water Works Assoc. 1984, 76, 65. (35) Rennecker, J. L.; Kim, J.-H.; Corona-Vasquez, B.; Marinas, B. Environ. Sci. Technol. 2001, 36, 2752. (36) Taylor, R. H.; Falkinham, J. O.; Norton, C. D.; LeChevallier, M. W. Appl. Environ. Microbiol. 2000, 66, 1702. (37) Narkis, N.; Katz, A.; Orshansky, F.; Kott, Y.; Friedland, Y. Water Sci. Technol. 1995, 31, 105. (38) Willberg, D. M.; Lang, P. S.; Hochemer, R. H.; Kratel, A.; Hoffmann, M. R. Environ. Sci. Technol. 1996, 30, 2526. (39) Willberg, D. M.; Lang, P. S.; Hochemer, R. H.; Kratel, A.; Hoffmann, M. R. CHEMTECH 1996, 26 (4), 53. (40) Hochemer, R. Ph.D. Thesis, California Institute of Technology, 1996. (41) Lang, P. S.; Ching, W.-K.; Willberg, D. M.; Hoffmann, M. R. Environ. Sci. Technol. 1998, 32, 3142. (42) Buntzen, R. R. The Use of Exploding Wires in the Study of SmallScale Underwater Explosions. In Exploding Wires, Vol. 2; Chace, W. G., Moore, H. K., Eds.; Plenum Press: New York, 1962; p 195. (43) Smith, K. F. Electro-Hydraulic Forming. In High-Velocity Forming of Metals; Wilson, F. W., Ed.; Prentice Hall: Englewood Cliffs, NJ, 1964; p 77. (44) Wesley, R. H.; Ayres, R. A. U.S. Patent No. 4,479,680, 1984. (45) Howard, D. D. Ph.D. Thesis, California Institute of Technology, 1996. (46) Howard, D.; Sturtevant, B. Ultrasound Med. Biol. 1997, 23, 1107. (47) Lokhanwalla, M.; Sturtevant, B. Phys. Med. Biol. 2000, 45, 1923. (48) Kratel, A. W. H. Ph.D. Thesis, California Institute of Technology, 1996. (49) Edebo, L.; Selin, I. J. Gen. Microbiol. 1968, 50, 253. (50) Gilliland, S. E.; Speck, M. L. Appl. Microbiol. 1967, 15, 1031. (51) Gilliland, S. E.; Speck, M. L. Appl. Microbiol. 1967, 15, 1038. VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4143

(52) Robinson, J. W.; Ham, M.; Balaster, A. N. J. Appl. Phys. 1973, 44, 72. (53) Miller, J. H. Experiments in Molecular Genetics, 3rd ed.; Cold Springs Harbor: Cold Springs Harbor, NY, 1972. (54) Pardee, A. B.; Jacob, F.; Monod, J. J. Mol. Biol. 1959, 1, 165. (55) Loomis, T. A. Numbers in Toxicology. Essentials in Toxicology; Lea & Febiger: Philadelphia, 1978; p 13. (56) Cerf, O. J. Appl. Bacteriol. 1977, 42, 1. (57) Johnston, M. D.; Simons, E. A.; Lambert, R. J. W. J. Appl. Microbiol. 2000, 88, 237. (58) Lambert, R. J. W.; Johnston, M. D. J. Appl. Microbiol. 2000, 88, 907. (59) Parker, J. A.; Darby, J. L. Water Environ. Res. 1995, 67, 1065. (60) Qualls, R. G.; Flynn, M. P.; Johnson, J. D. J. Water Pollut. Control Fed. 1983, 55, 1280. (61) Cerf, O.; Davey, K. R.; Sadoudi, A. K. Food Res. Int. 1996, 29, 219. (62) Henglein, A. Contributions to Various Aspects of Cavitation Chemistry. In Advances in Sonochemistry; Mason, T. J., Ed.; JAI Press: Greenwich, CT, 1993; p 17. (63) Henglein, A.; Gutierrez, M. J. Phys. Chem. 1990, 94, 5169. (64) Nikogosyan, D. N.; Letokhov, V. S. Riv. Nuovo Cimento 1983, 6, 1.

4144

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001

(65) Nikogosyan, D. N.; Angelov, D. A.; Oraevsky, A. A. Photochem. Photobiol. 1982, 35, 627. (66) Schulte-Frohlinde, D.; Simic, M. G.; Gorner, H. Photochem. Photobiol. 1990, 52, 1137. (67) Masnyk, T. W.; Minton, K. W. Photochem. Photobiol. 1991, 54, 99. (68) Kochevar, I. E.; Walsh, A. A.; Green, H. A.; Sherwood, M.; Shih, A. G.; Sutherland, B. M. Cancer Res. 1991, 51, 288. (69) Schulte-Frohlinde, D.; Bothe, E. Pulse Radiolysis of Nucleic Acids in Aqueous Solutions; CRC Press: Boca Raton, FL, 1991. (70) Angelov, D.; Berger, M.; Cadet, J.; Getoff, N.; Keskinova, E.; Solar, S. Radiat. Phys. Chem. 1991, 37, 717. (71) Ching, W.-K.; Colussi, A. J.; Hoffmann, M. R. Environ. Sci. Technol. Manuscript in preparation.

Received for review February 14, 2001. Revised manuscript received July 31, 2001. Accepted July 31, 2001. ES010643U