Environ. Sci. Technol. 2001, 35, 4103-4110
Microbial Inhibitors for U.S. EPA Drinking Water Methods for the Determination of Organic Compounds STEPHEN D. WINSLOW,* BARRY V. PEPICH, MARGARITA V. BASSETT, AND STEVEN C. WENDELKEN IT Corporation, 26 West Martin Luther King Drive, Cincinnati, Ohio 45268 DAVID J. MUNCH AND JAMES L. SINCLAIR United States Environmental Protection Agency, Office of Groundwater and Drinking Water, 26 West Martin Luther King Drive, Cincinnati, Ohio 45268
Preservation of chemical analytes in drinking water samples is necessary to obtain accurate information concerning contaminant occurrence. Sample preservation to prevent biodegradation is important for most samples and analytes. With the unique demands of environmental methods, it is not always possible to kill all microorganisms without having undesirable effects. To find a suitable preservative, the sample, analysis, and preservation needs should be considered. During method development of U.S. Environmental Protection Agency (EPA) Methods 526 (for unstable semivolatile compounds) and 532 (for phenylurea pesticides), a number of studies were conducted to identify compatible microbial inhibitors. Copper sulfate was successfully used in Method 532 and is an excellent firstchoice antimicrobial agent for many applications. Copper sulfate can catalyze hydrolysis reactions for some pesticides such as those analyzed in Method 526. Under these conditions, a nonmetal compound of antimicrobial activity must be considered. During the development of Method 526, a survey of alternate organic based antimicrobial compounds found that diazolidinyl urea worked well in the method. Several other candidate microbial inhibitors were identified that could have application to other environmental methods. A general approach to selecting antimicrobial compounds in future environmental methods in water matrixes is discussed.
Introduction Preservation of chemical analytes in drinking water samples is a necessary part of obtaining accurate contaminant occurrence information. Analytes in samples may be lost through volatilization, sorption, abiotic chemical reactions, and microbial degradation. Microorganisms have the potential to degrade target analytes and represent a significant pathway for the fate and destruction of organic compounds (1). Microbial degradation of organic compounds is influ* Corresponding author phone: (513)569-7035; fax: (513)569-7837; e-mail:
[email protected]. 10.1021/es010661b CCC: $20.00 Published on Web 09/12/2001
2001 American Chemical Society
enced by a number of factors. These factors include chemical composition (2) and the concentration of target analytes (3). Characteristics of the water sample may also affect biodegradation of analytes, such as population sizes of specific degraders (4), nutrients (5), and other biodegradable organic compounds that may increase (6) or decrease (7) analyte degradation. Therefore, sample preservation to prevent biodegradation of analytes is important for most samples and analytes. Both untreated source water and finished drinking water samples have been shown to require preservation. Analytes in ambient water have been lost due to microbial action (8). Microbial regrowth has been reported in municipal waters that have been dechlorinated and that contain assimilable organic carbon (9). Dechlorination of a chlorinated water sample is a required step in most drinking water methods because residual chlorine may degrade target analytes. Once an analytical sample is dechlorinated, an antimicrobial agent is needed. A number of sample preservation and storage stability studies have been conducted in an effort to extend laboratory sample holding times without jeopardizing data quality (7, 10-12). The goal of preservation is to decrease the rate of target analyte degradation to an acceptable level. Antimicrobial preservation prevents biologic deterioration of materials by suppressing organisms that may be present in the sample (13). The effectiveness of an antimicrobial agent will be different for Gram-negative bacteria, Gram-positive bacteria, fungi, and molds. The minimum bactericidal concentration will also be different for each microorganism. With the unique demands of environmental methods, it is not always possible to kill all organisms without having undesirable effects. In the past, many U.S. Environmental Protection Agency (EPA) drinking water methods used acid preservation to pH less than 2 or mercuric chloride to minimize microbial growth in water samples. Low solution pH can increase the degradation rates of compounds subject to acid-catalyzed hydrolysis. Data from the Agricultural Research Service indicate that hydrolysis rates of cyanazine, diazinon, and disulfoton are highest at low pH (14). Work in this laboratory confirmed higher hydrolysis rates at low pH for these compounds (15). Mercuric chloride can also increase degradation rates of compounds subject to acid-catalyzed hydrolysis (15-17). During the initial 14-day storage stability study of the U.S. EPA National Pesticide Survey, 26 of 147 target analytes had 100% loss in recovery, which was probably due, in some cases, to preservation with mercuric chloride (18). Other negative aspects of mercury compounds are high human toxicity and the expense of hazardous waste disposal. An ideal antimicrobial agent should have several general characteristics. It must inhibit microbial growth in the presence of assimilable organic carbon and keep microbial populations low under method conditions. It must be stable during storage and inert to target analytes and other preservative compounds. The antimicrobial agent should not precipitate with elements of the drinking water matrix. It must have high water solubility and should be minimally retained by the extraction solvent or solid-phase extraction (SPE) material so that it does not interfere with the quantitation of analytes, surrogates, and internal standards. It should not cause additional instrument maintenance, such as cleaning of the detector source or replacement of the capillary column or the injection port liner. And finally, an ideal microbial inhibitor should be relatively safe to handle, reasonably priced, and allow nonhazardous disposal. VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4103
Over the past 3 years, our laboratory has conducted a number of studies to identify microbial inhibitors for a variety of analytes of environmental concern in raw and finished waters. This paper summarizes the approach we used to select and test antimicrobial agents during the development of two recently published EPA drinking water methods, Methods 526 (12) and 532 (11). For Method 532 where the compounds possess good stability, the selection of a suitable antimicrobial agent was relatively straightforward. The task was much more challenging for Method 526, which has labile target compounds that are easily degraded in solution. Through this work, we identified several antimicrobial agents that were not employed in the preservation schemes of our published methods but are strong candidates for future environmental methods. Finally, we describe a general approach to selecting microbial compounds in future methods.
Experimental Section Chemical Reagents. The following antimicrobial compounds were investigated: benzalkonium chloride, benzyldimethylhexadecylammonium chloride, cetyl pyridinium chloride, diazolidinyl urea, and trisodium ethylenediaminetetraacetic acid (Sigma, St. Louis, MO); copper sulfate and mercuric chloride (ACS grade, Fisher Scientific, Pittsburgh, PA); chlorhexidine diacetate, 1-hydroxypyridine-2-thione zinc salt, imidazolidinylurea, isonicotinic acid hydrazide free base, methyl paraben, and propyl paraben (ICN, Aurora, Oh); kathon (a solution of 1.15% 5-chloro-2-methyl-4-isothiazolin3-one and 0.35% 2-methyl-4-isothiazolin-3-one in water, Supelco, Bellefonte, PA); sodium omadine (40% aqueous, Arch Chemical, Norwalk, CN); 5-chloro-2-(2,4-dichlorophenoxy)phenol (Ciba-Geigy, Greensboro, NC); and 2-chloroacetamide, 2-mercaptopyridine N-oxide sodium salt hydrate, methylene dithiocyanate, phenethyl alcohol, sodium thiocyanate, 3,3,4′-trichlorocarbanilide, and tris(hydroxymethyl)nitromethane (Aldrich, Milwaukee, WI). The growth media for the heterotrophic plate studies was R2A Agar (Difco Laboratories, Detroit, MI). Sample Preparation and Analysis. A complete description of instrumental conditions, chemical reagents, and analytical procedures can be found elsewhere (11, 12). For Method 526, a 1-L water sample containing preservatives was extracted by styrene divinylbenzene SPE cartridge or disk (Varian, Inc., Palo Alto, CA) and concentrated to a final volume of 1 mL. The extracts were analyzed using a GC/MSD (model 6890/5973, Hewlett-Packard Co., Wilmington, DE) equipped with a fused silica capillary column (30 m × 0.25 mm i.d., 0.25 um film, DB-5MS, J&W Scientific, Inc., Folsom, CA). For Method 532, a 500-mL water sample containing preservatives was extracted with a C18 SPE cartridge or disk (Varian, Inc., Palo Alto, CA) and concentrated to a final volume of 1 mL. Extracts were analyzed using a high-performance liquid chromatography (HPLC) system (Waters model 2690) equipped with a photodiode array detector (Waters model 996) and a Waters symmetry column (4.6 × 150 mm, packed with 3.5 µm dp C18 stationary phase, Waters Chromatography Division, Milford, MA). Instrument conditions and extraction procedures are fully described in EPA Methods 526 and 532 (11, 12). Storage Stability Studies. Storage stability studies were designed to track percent recovery of target analytes over time. Separate studies were run to examine the antimicrobial effectiveness and method compatibility of preservation agents. Some target analyte compounds are not readily consumed by the general microbial population but could be degraded by a population of microbes that have been acclimated to the compound by repeated exposure to the chemical (19). Since an acclimated microbe population of specific chemical degraders was not available for our studies, 4104
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001
the effectiveness of a microbial inhibitor was evaluated by enumerating culturable microorganisms detected with a heterotrophic plate count method (20). Colony forming units (CFUs) per milliliter of sample were used as an indication of viable microorganisms, although the numbers of the heterotrophic microorganisms detected with this method may not reflect populations of specific chemical degraders (21). In these experiments, sample bottle preparation and sample collection and storage were conducted in the same manner as real samples. Since drinking water samples have very small microbial populations, an aliquot (1-10 mL) of raw water containing indigenous microorganisms was added to a subset of samples (a volume of 1 L for Method 526 and of 500 mL for Method 532) to challenge the effectiveness of the antimicrobial compounds. Sufficient raw water was added to give an HPC study between 50 and 200 counts per plate for control samples. Pour plate studies of antimicrobial compounds were performed on R2A Agar inoculated with 0.20 mL of sample, incubated at 35 °C. CFUs were counted 5 days after inoculation. Storage stability studies were conducted at the maximum temperatures allowed by Methods 526 and 532 for shipping and storage, that is, at 10 °C for the first 48 h and at 6 °C for the remainder of storage. The initial temperature at 10 °C was chosen as a maximum permissible temperature when samples are shipped in boxes cooled using chemical freezer packs. Antimicrobial Screening Studies. Screening of antimicrobial compounds was sometimes performed at room temperature to hasten method development, but final evaluation of microbial inhibitors was always performed under actual method storage conditions. At lower temperature, the antimicrobial agent will be less effective in suppressing microbial growth because the rate of chemical reactions of microbial agents will decrease at lower temperature, while certain microbe populations can acclimate to a lower temperature. Successful preservation of analytes in a stability study at 20 °C does not ensure that it will work well at 6 °C. For example, in our lab, sodium thiocyanate suppressed growth at 20 °C for 7 days, but at 6 °C after 3 days, the microbe population showed exponential growth. Gelinas et al. (22) reported the comparative effect of temperature on disinfectants at minimal bactericidal concentration, which showed that more disinfectant was required with decreasing temperature. Screening of antimicrobial compounds was performed on spread plates of R2A Agar that were inoculated with 0.10 mL of sample, incubated at 35 °C, and CFUs were counted 2 days after inoculation.
Results and Discussion Antimicrobial Families. Antimicrobial compounds can be categorized into groups based on chemical structure (23). These chemical families are summarized in Table 1 with their modes of action. The applications of specific members of these families are described in the two sections that follow. Several antimicrobial families, including oxidants, surface active agents, and organotin compounds, were not investigated during the development of Methods 526 and 532. The reactivity of oxidants prevents their use as antimicrobial agents in many EPA methods. Organotin compounds were not considered because those that have low human toxicity have very low water solubility. Surface active agents (surfactants) are potential antimicrobial agents for drinking water methods despite their potential disadvantages. Included in this group are quaternary ammonium salts (QAS), acid anionic compounds, and amphoteric compounds. Most surfactants cause water solutions to foam during the extraction process and are retained
TABLE 1. Antimicrobial Families and Proposed Mode of Action antimicrobial familiesa strong oxidizing agents: chlorine and chlorine compds, iodine and iodine compds, peroxygen compds, and ozone alcohols
proposed mode of actiona
application
oxidizer combines with components of cell protoplasm, destroying organism; one theory suggests chlorine combines with proteins of cell membranes, forming N-chloro compds that interfere with cell metabolism denatures proteins and inhibits production of metabolites essential for cell division
not investigated
limited investigation in Method 526 phenolic compds at high concentration, gross protoplasmic poison and precipitator limited of cell proteins; at lower concentrations inactivates essential investigation enzyme systems in Method 526 quaternary ammonium salts (QAS) adsorption on bacterial cell surface, diffusion through cell wall, investigated in binding to cytoplasmic membrane, causing disruption of the Method 526 cytoplasmic membrane and release of K+ ions and other cytoplasmic constituents followed by precipitation of cell contents and death of cells surface-active agents (surfactants) acid anionic compds promote disorganization of cell membrane not investigated and inhibit key enzymic activities; interruption of cellular transport and denaturation of cellular proteins; amphoteric compds cause protein precipitation nitrogen compds (such as sodium azide, simple nitrogen compounds, e.g., sodium azide, tie up heme iron; investigated in formaldehyde condensates, nitrites, formaldehyde condensates release formaldehyde which Method 526 nitriles, nitro derivatives, pyridines, alkylates nucleic acids; nitrite action can inhibit of germination thiazoles, and amine derivatives) and cell division; nitriles and nitro derivatives inhibit oxidative phosphorylation; pyridines of type nicotinamide interfere with synthesis of long-chain fatty acids and with transaminase activity; thiazoles block electron transport; amine derivatives have mostly unspecified modes of action such as lowering surface tension and adding to alkaline reserve polymeric antimicrobial agents overwhelms bacterial cell wall exclusion mechanisms to promote limited leakage of low molecular weight cytoplasmic components and investigation inhibits certain membrane-bound enzymes; precipitates in Method 526 cytoplasm by formation of complexes with phosphated entities mercury salts combines with thiol groups of enzymes investigated in Method 526 solution pH extremes of pH can effectively inhibit microbial growth; viability investigated in for most microbes is limited to the pH range of 4.5-9 Method 526 organotin compds metabolic inhibitors that interfere with microbial respiration not investigated copper and zinc preservatives combines with thiol groups of enzymes investigated in Methods 526 and 532 a
Information summarized from ref 23.
by the SPE material. For these reasons, only QASs were evaluated during these studies. Selection of an Antimicrobial Inhibitor for EPA Method 532. Copper sulfate, sodium omadine, and kathon were selected as initial antimicrobial candidates for EPA Method 532. Copper sulfate has been used to inhibit microbial growth for many years (24). It has high water solubility and is nonhazardous for handling and disposal. The antimicrobial action of copper sulfate is believed to derive from the ability of the copper ions, which have penetrated the cell wall, to chelate thiol groups, thereby interfering with the cell proteins or enzymes and poisoning the cell (13, 25). Sodium omadine and kathon are members of the nitrogen family, and their modes of action are summarized in Table 1 under the pyridines and thiazoles, respectively. Like copper sulfate, they have high water solubility. A preliminary test of antimicrobial effectiveness of copper sulfate (500 mg/L), sodium omadine (64 mg/L), and kathon (15 mg/L) was performed using reagent water inoculated with an aliquot of creek water (1 mL/L). Additionally, a control set was prepared in an identical manner without the preservative. The samples were allowed to sit at room temperature for 1 day and then were pour plated. CFUs were counted 5 days after incubation at 35 °C. These results are summarized in Table 2. To evaluate the compatibility of these reagents with the Method 532 compounds, samples were fortified with Method 532 analytes at 10 µg/L in reagent water, stored at 4 °C, and
TABLE 2. Initial Screening of Antimicrobial Effectiveness after 1 Day of Storage for Copper Sulfate, Sodium Omadine, and Kathon antimicrobial agent
concn of antimicrobial agent (mg/L)
unpreserved control samples (av CFU/mL)
preserved samples (av CFU/mL)
copper sulfate sodium omadine kathon
500 64 15
4000 8900 4900
1 112 29
analyzed in triplicate on days 0, 7, and 15. Day 0 results for the sodium omadine preserved samples had a large early eluting interferant from residual sodium omadine that was retained by the C18 solid phase. The interferant prevented the quantitation of tebuthiuron, yielding a 160% average recovery, which rendered sodium omadine unsuitable for Method 532. Copper sulfate and kathon reagent water sample recoveries were acceptable for all compounds (see Figure 1). On the basis of these initial results, a 21-day storage stability study in chlorinated surface water was initiated for copper sulfate and kathon. Experimental conditions were identical to those described above, but ammonium chloride was added (100 mg/L) to sequester free chlorine. The samples containing copper sulfate and ammonium chloride had acceptable recoveries over the 21-day study for all compounds and demonstrated excellent antimicrobial activity (26). For VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4105
FIGURE 1. Percent recovery for Method 532 target analytes in reagent water containing copper sulfate at 500 mg/L and kathon at 15 mg/L. Error bars represent ( 1σ.
TABLE 3. Microbial Inhibitors That Were Eliminated from Consideration Based on Physical Properties compound
family
benzalkonium chloride benzyldimethylhexadecylammonium chloride cetylpyridinium chloride chlorhexidine diacetate
QASa
5-chloro-2-(2,4-dichlorophenoxy)phenol methyl paraben propyl paraben 1-hydroxypyridine-2-thione, zinc salt 2-mercaptopyridine N-oxide, sodium salt hydrate methylene dithiocyanate phenethyl alcohol 3,3,4′-trichlorocarbanilide tris(hydroxymethyl)nitromethane
phenol phenol phenol nitrogen nitrogen nitrogen alcohol nitrogen nitrogen
a
QAS, quaternary ammonium salt.
b
QASa QASa PAAb
9
foamed during extraction and retained on SDVB SPE material foamed during extraction and retained on SDVB SPE material foamed during extraction and retained on SDVB SPE material precipitates when added to hard water containing inorganic anions such as sulfate and carbonate limited water solubility retained on SDVB SPE material limited water solubility very limited water solubility retained on SDVB SPE material retained on SDVB SPE material highly retained on SDVB SPE material, odiferous limited water solubility interfered with GC/MS quantitation of 2,4,6-trichlorophenol
PAA, polymeric antimicrobial agent.
the samples that contained kathon and ammonium chloride, day 0 recoveries for thidiazuron, propanil, siduron A and B, and diflubenzuron were all below 50%, indicating an adverse reaction associated with this combination of reagents. The use of kathon was not investigated further because it was not available in solid form, which is preferable to simplify shipment of sample bottles containing preservatives to sampling site. It should be noted that recoveries in kathon samples were acceptable in the reagent water without ammonium chloride so that kathon may still have use as an antimicrobial compound for water matrixes that do not require a dechlorinating agent. In subsequent experiments, copper sulfate was found to form a blue precipitate in a tested groundwater matrix at neutral pH. The use of Tris buffer was found to prevent the formation of precipitate. Tris may be acting as a bulky ligand that sterically excludes the precipitating counterion. Fitzgerald (27) reported that chelated copper stayed in solution better than nonchelated copper. The use of Tris with the phenylurea pesticides of Method 532 had the additional advantage of sequestering free chlorine, possibly as chloramines, thereby eliminating the need for ammonium chloride to protect the phenylureas. The final 4106
limitations
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001
Method 532 conditions required the addition of copper sulfate as the antimicrobial agent and a Tris salt mixture that acted as a pH 7 buffer and as the dechlorinating agent and prevented precipitation of copper sulfate in hard water. Analyses using Method 532 demonstrated excellent precision and accuracy in surface waters and groundwaters (11). Selection of an Antimicrobial Inhibitor for EPA Method 526. Initial method development studies found that preservation by the use of acidic conditions and by the addition of copper sulfate or mercuric chloride was unacceptable. Sample storage at pH less than 2 caused catalytic hydrolysis of 1,2-diphenylhydrazine, terbufos, diazinon, disulfoton, and cyanazine. Copper sulfate, even at low concentrations near the regulatory limit for copper set by the EPA, catalyzed hydrolysis of cyanazine, terbufos, diazinon, and disulfoton (12). Mercuric chloride reduced target analyte recoveries due to catalytic hydrolysis as well. As a result, a search for alternative, nonmetal microbial inhibitors was made. An initial screening of antimicrobial compounds indicated that 13 potential candidates could not be used based on their physical properties, as summarized in Table 3. Many compounds in the family of nitrogen-containing microbial
FIGURE 2. Percent recoveries of Method 526 analytes for 21-day stability study with samples containing sodium omadine at 64 mg/L. All samples contain preservation agents trisodium EDTA (0.35 g/L), Tris buffer at pH 7 (7.80 g/L total), and ascorbic acid (0.10 g/L). Error bars represent ( 1σ.
FIGURE 3. Heterotrophic plate count results for 21-day storage stability study of antimicrobial effectiveness of sodium omadine at 64 mg/L in drinking water from a surface water utility. The concentration of the reagents specified are 0.35 g/L for trisodium EDTA, 7.80 g/L total for Tris buffer (pH 7), and 0.10 g/L for ascorbic acid. “Seeded” means an aliquot of raw water was added to the sample. inhibitors had potentially broad application to drinking water methods because of their high water solubility. Sodium omadine was the first compound in the nitrogen family to be studied. Figure 2 shows percent recoveries for Method 526 analytes during a 21-day storage stability study performed in drinking water from a surface water utility using sodium omadine as a microbial inhibitor. Even at 21 days, all recoveries were above 70%. Results from a heterotrophic plate count (HPC) study covering a 21-day period, shown in Figure 3, indicated that sodium omadine repressed microbe
population growth effectively at a concentration of 64 µg/ mL. In a precision and accuracy study in reagent water with sodium omadine, shown in Table 4, analytes fortified at a low level (0.5 µg/L) were recovered between 72.6 and 114%, which is acceptable in the method for a level equivalent to the lowest calibration standard. At a high fortification, 20 µg/L, in reagent water with sodium omadine, recoveries were between 84.4 and 102%. Unfortunately, when fortified groundwater samples (400 mg/L in equiv CaCO3) were extracted, a green precipitate formed in the extracts. The VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4107
TABLE 4. Percent Recovery for Low and High Spikes in Reagent Water and in Drinking Water from a Groundwater Utility Containing Sodium Omadine at 64 mg/La reagent water (0.50 µg/L)
nitrobenzene 2,4-dichlorophenol 2,4,6-trichlorophenol 1,2-diphenylhydrazine prometon terbufos dyphonate diazinon disulfoton acetochlor cyanazine a
reagent water (20 µg/L)
groundwater (0.50 µg/L)
groundwater (20 µg/L)
% rec
% RSD
% rec
% RSD
% rec
% RSD
% rec
% RSD
82.6 95.1 114 89.1 91.4 85.7 82.9 82.9 72.6 84.9 110
5.5 5.3 6.7 4.3 5.5 4.9 4.6 5.7 4.1 4.4 5.7
84.4 91.7 94.2 92.4 92.8 90.6 87.9 86.9 87.9 88.0 102
7.4 3.5 2.5 3.1 2.2 3.5 2.3 2.2 2.2 3.5 3.1
77.1 86.3 104 79.4 89.7 83.1 76.9 77.1 69.1 79.4 111
10.8 8.7 7.4 8.1 4.5 6.0 4.5 8.4 5.0 6.3 8.6
90.0 88.7 91.8 92.0 90.7 93.1 91.4 87.0 88.8 86.6 102
4.6 4.4 4.8 4.4 4.9 5.1 4.5 4.6 4.5 4.4 4.6
All samples contain preservation agents trisodium EDTA (0.35 g/L), Tris buffer at pH 7 (7.80 g/L total), and ascorbic acid (0.10 g/L).
FIGURE 4. Room temperature, 7-day storage stability studies for trisnitro (1000 mg/L), methylene bisthiocyanate (MBT) (500 mg/L), 2-CA (2000 mg/L), sodium thiocyanate (2000 mg/L), and DZU (1000 mg/L). Samples of 1 L of drinking water from a groundwater utility (400 mg/L of equiv CaCO3) were fortified with Method 526 analytes at 5 µg/L. All samples contain preservation agents trisodium EDTA (0.35 g/L), Tris buffer at pH 7 (7.80 g/L total), and ascorbic acid (0.10 g/L). Error bars represent ( 1σ. percent recoveries at the low level were, on average, about 5% lower than in reagent water, and high-level fortification recoveries were about the same as reagent water. These recoveries were acceptable, but the presence of the precipitate had the potential to complicate extract manipulation and to possibly affect analyte recovery adversely. Kathon was the next microbial inhibitor to be tested. When ascorbic acid, the dechlorinating agent of Method 526, was mixed to samples bottle with kathon, a yellow color appeared, so the use of kathon was not pursued further. Seven additional antimicrobial nitrogen-containing compounds with high water solubility were tested to determine their applicability to Method 526: isonicotinic acid hydrazide, 4108
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001
phenethyl alcohol, tris(hydroxymethyl)nitromethane (trisnitro), methylene bisthiocyanate (MBT), 2-chloroacetamide, sodium thiocyanate, and diazolidinyl urea. When isonicotinic acid hydrazide was tested on day 0, cyanazine had only half the usual recovery, and phenethyl alcohol caused low recoveries for nitrobenzene (34.1%) and 2,4,6-trichlorophenol (69.3%). Due to these effects, isonicotinic acid hydrazide and phenethyl alcohol were removed from consideration. A 7-day screening study at room temperature was performed for the five remaining compounds. Results are shown in Figure 4. The samples with methylene bisthiocyanate had depressed recoveries overall, and so methylene bisthiocyanate was not investigated further. Trisnitro inter-
FIGURE 5. Recoveries at day 14 under EPA Method 526 storage conditions for DZU, 2-CA, and sodium thiocyanate. Analytes were fortified at 5 µg/L in 1 L of drinking water from a surface water utility. All samples contain trisodium EDTA (0.35 g), Tris buffer at pH 7 (7.80 g total), and ascorbic acid (0.10 g). Error bars represent ( 1σ.
TABLE 5. Heterotrophic Plate Count Data for 21 Days under Method Storage Conditions for DZU, 2-CA, Sodium Thiocyanate, and Control Samplea day
DZU (av CFU/mL)
2-CA (av CFU/mL)
sodium thiocyanate (av CFU/mL)
control (av CFU/mL)
0 3 7 14 21
79 15 25 30 23
102 36 104 152 53
76 24 2 300 200 000 180 000
111 51 3 100 220 000 140 000
a All samples contain trisodium EDTA (0.35 g), Tris buffer at pH 7 (7.80 g total), and ascorbic acid (0.10 g). Control sample does not contain a microbial inhibitor.
fered with quantitation of 2,4,6-trichlorophenol (m/z 196). Trisnitro did not interfere with other analytes and might have application in another method. On the basis of these results, a 21-day storage stability study was initiated using sodium thiocyanate, 2-chloroacetamide (2-CA), and diazolidinyl urea (DZU) under Method 526 final conditions. Analyte recovery and HPCs were determined. Figure 5 presents data through day 14 of this study. Most analyte recoveries in sodium thiocyanate, 2-CA, and DZU solutions were similar, except that terbufos in sodium thiocyanate solution recovered 20% less than either of the other two, and disulfoton recovery for 2-CA was 10% less. Overall, samples containing DZU showed slightly higher analyte recoveries than those containing sodium thiocyanate and 2-CA. Results of a HPC study comparing sodium thiocyanate, 2-CA, and DZU are shown in Table 5. DZU and 2-CA had relatively low CFUs, which indicated a repression of microbial growth. Sodium thiocyanate samples showed strong microbial growth, indicating that sodium thiocyanate was not effective as a microbial inhibitor. DZU was chosen over 2-CA as the microbial inhibitor for Method 526 because it repressed microbial growth better than 2-CA, did not have the health
effect warnings of 2-CA, and dissolved much faster in water than 2-CA. Diazolidinyl urea was originally proposed in 1982 as a water-soluble, broad spectrum, microbial inhibitor for cosmetic use and is presently in use in products such as suntan lotion (28). Antimicrobial activity was proposed to be due to protein alkylation of sulfhydryl groups and the ability to release formaldehyde (29). The minimum inhibitory concentration for DZU was reported as 250 mg/L for Grampositive bacteria and 1000 mg/L for Gram-negative bacteria (30). General Approach in Selecting Antimicrobial Compounds for Environmental Methods in Water Matrixes. Several options are available for preserving water samples from microbiological degradation during storage. If low pH does not increase degradation of target analytes or interfere with extraction, adjustment of pH to below 2 remains a costeffective antimicrobial treatment. If the use of a strong acid at the sampling site must be avoided or if the analytes require storage at neutral pH, copper sulfate should be the next choice. Copper sulfate is an effective antimicrobial agent that can be kept in solution even in hard water matrixes using a Tris buffer. In chlorine-disinfected waters, Tris buffer sequesters free chlorine. If target analytes and preservation agents tolerate this condition, then Tris may function as buffer and dechlorinating agent, and the addition of an additional dechlorinating agent may not be necessary. If copper ion catalyzes hydrolysis of compounds, a nonmetal antimicrobial agent, such as DZU, should be considered. While DZU was found to be most compatible with Method 526, our studies indicated that 2-CA and sodium omadine were also promising candidates for the antimicrobial preservation of labile compounds. Compounds in the family of nitrogen-containing microbial inhibitors will have broad application to drinking water methods because of their high water solubility. When selecting new microbial inhibitors, it is most costeffective to initially screen these candidates in the presence VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4109
of all other preservation agents (buffers, dechlorinating reagents, etc.). To ensure chemical compatibility, method extraction and analyses should be performed with all preservatives and target analytes. Even though target analytes may not microbially degrade under laboratory conditions, their stability in field samples cannot be assumed. Therefore, antimicrobial action of the preservation scheme should be confirmed under sample storage conditions using a procedure similar to the HPC studies described above. Precision and accuracy determinations should be made with the complete set of selected preservatives. If precision and accuracy determinations are satisfactory, a final evaluation should be made in a range of matrixes that span the intended scope or applicability of the method. For drinking water methods, in lieu of testing many waters, a hard water (subsurface source containing ∼400 mg/L equiv CaCO3) and a surface water should be used. The subsurface water will challenge the method with a high carbonate matrix that contains metal ions, and the surface water matrix will test the method in terms of high total organic carbon. In our studies, groundwater presented the greatest challenges for both the inorganic based and the organic based preservation schemes. This general approach offers a means to identify antimicrobial agents for the analysis of compounds of environmental concern in water matrixes. The application of a suitable preservative for an analytical method can be relatively uncomplicated, but systematic and careful testing is required to demonstrate robust method performance in water matrixes of intended use.
(8) (9) (10) (11)
(12)
(13) (14)
(15) (16)
(17) (18) (19) (20)
Acknowledgments All work was supported on-site at the EPA’s Drinking Water Laboratory located in Cincinnati, OH. This work has been funded wholly or in part by the United States Environmental Protection Agency under an on-site contract (Contract 68C6-0040) to IT Corporation. It has been subject to the Agency’s review, and it has been approved for publication as an EPA document. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
Literature Cited (1) Barbash, J. E.; Resek, E. A. Pesticides in Ground Water; Pesticides in the Hydrologic Series; Ann Arbor Press: Chelsea, MI, 1996; Vol. 2. (2) Boethling, R. S.; Sabljic, A. Environ. Sci. Technol. 1989, 23 (6), 672. (3) Boethling, R. S.; Alexander, M. Appl. Environ. Microbiol. 1979, 37 (6), 1211. (4) Ventullo, R. M.; Larson, R. J. Appl. Environ. Microbiol. 1986, 51 (2), 356. (5) Swindoll, C. M.; Aelion, C. M.; Pfaender, F. K. Appl. Environ. Microbiol. 1988, 54 (1), 212. (6) Schmidt, S. K.; Alexander, M. Appl. Environ. Microbiol. 1985, 49 (4), 822. (7) Munch, J. W.; Munch, D. J.; Winslow, S. D.; Wendelken, S. C.; Pepich, B. V. U.S. EPA Method 556, Revision 1.0; National
4110
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001
(21) (22) (23) (24) (25) (26)
(27) (28) (29) (30)
Exposure Research Laboratory, Office of Research and Development, U.S. EPA: Cincinnati, OH, 1998. Shukairy, H. M. The Control of Disinfection By-Product Formation by Ozonation and Biotreatment. Ph.D. Dissertation, University of Cincinnati, 1994. Power, K. N.; Nagy, L. A. Water Res. 1999, 33 (3), 741. Munch, D. J.; Munch, J. W.; Pawlecki, A. M. EPA Method 552.2, Revision 1.0; National Exposure Research Laboratory, Office of Research and Development, U.S. EPA: Cincinnati, OH, 1995. Bassett, M. V.; Wendelken, S. C.; Dattilio, T. A.; Pepich, B. V.; Munch, D. J. EPA Method 532, Revision 1.0; National Exposure Research Laboratory, Office of Research and Development, U.S. EPA: Cincinnati, OH, 2000. Website: http://www.epa.gov/ ogwdw000/methods/532.pdf. Winslow, S. D.; Prakash, B.; Domino, M. M.; Pepich, B. V.; Munch, D. J. EPA Method 526; National Exposure Research Laboratory, Office of Research and Development, U.S. EPA: Cincinnati, OH, 2000. Website: http://www.epa.gov/ogwdw000/methods/ 526.pdf. Russel, A. D. Principles of Antimicrobial Activity. In Disinfection, Sterilization, and Preservation, 4th ed.; Block, S. S., Ed.; Lea & Febiger: Philadelphia, 1991; pp 29-58. Agricultural Research Service Pesticide Properties Database. U.S. Department of Agriculture, Agricultural Research Service, Beltsville, MD, February 1, 2001. Website: http://www.arsusda.gov/rsml/ppdb.html. Winslow, S. D.; Prakash, B.; Domino, M. M.; Pepich, B. V.; Munch, D. J. Environ. Sci. Technol. 2001, 35 (9), 1851. Smolen, J. M.; Stone, A. T. Preprint of Papers Presented at the 208th ACS National Meeting, Washington, DC., August 21-25, 1994; Vol. 34, No. 2, Division of Environmental Chemistry; American Chemical Society: Washington, DC, 1994. Mabey, W.; Mill, T. J. Phys. Chem. Ref. Data 1978, 7 (2), 383. Munch, D. J.; Frebis, C. P. Environ. Sci. Technol. 1992, 26, 921. Racke, K. D.; Coates, J. R. J. Agric. Food Chem. 1987, 35, 94. Standard Methods 9215A and 9215B. In Standard Methods for the Examination of Water and Wastewater, 19th ed.; Eaton, A. D., Clesceri, L. S., Greenberg, A. E., Eds.; American Public Health Association; Washington, DC, 1995. Thomas, J. M.; Bruce, C. L.; Gordy, V. R.; Duston, K. L.; Hutchins, S. R.; Sinclair, J. L.; Ward, C. H. J. Ind. Microbiol. Biotechnol. 1997, 18, 313. Gelinas, P.; Goulet, J.; Tastayre, G. M.; Picard, G. A. J. Food Prot. 1984, 47 (11), 841. Block, S. S., Ed. Disinfection, Sterilization, and Preservation, 4th ed.; Lea & Febiger: Philadelphia, 1991. Hale, F. E. The Use of Copper Sulphate in Control of Microscopic Organisms; Phelps Dodge Refining Corp.: New York, 1954. Yeager, C. C. Copper and Zinc Preservatives. In Disinfection, Sterilization, and Preservation, 4th ed.; Block, S. S., Ed.; Lea & Febiger: Philadelphia, 1991; pp 358-361. Bassett, M. V.; Wendelken, S. C.; Dattilio, T. A.; Pepich, B. V.; Munch, D. J. The application of Trizma and copper sulfate for the preservation of phenylurea pesticides analyzed using U.S. EPA Method 532 in the UCMR survey. Manuscript in preparation for submission to Environ. Sci. Technol. Fitzgerald, G. P. Appl. Microbiol. 1960, 8, 269. Berke, P. A.; Rosen, W. E. Cosmet. Toiletries 1982, 97, 49-53. Llabres, C. M., Ahearn, D. G. Appl. Environ. Microbiol. 1985, 49 (2), 370. Kabara, J. J. Cosmetic and Drug Preservation; Marcel Dekker: New York, 1984.
Received for review February 20, 2001. Revised manuscript received July 23, 2001. Accepted July 31, 2001. ES010661B
