Journal of Applied Bacteriology 1991,71, 531-538
Decreased biocide susceptibility of adherent Legionella pneumophila J.B. Wright, 1. Ruseska and J.W. Costerton Department of Biological Sciences, University of Calgary, CalgaryvAlberta, Canada 3513/10/90:accepted 1 June 1991
a study of the in vitro effectiveness of biocides against Legionella pneumophila, some aspects of the cooling tower environment were replicated in the laboratory, paying particular attention to water hardness and pH. Pieces of Douglas fir and polyvinyl chloride were colonized in a recirculating system and the comparative efficacy of two biocides (Bronopol and Kathon) against the sessile and planktonic populations was examined. While the biocides were relatively effective against the planktonic L . pneumophila population over a short period of time (minimum 9-12 h), substantially longer periods of time (maximum > 48 h) were required to reduce the number of cultivable bacteria to below detectable levels in the adherent population. The results indicate that failure to monitor the sessile population of L. pneumophila in laboratory studies of biocides may result in the use of incorrect dosages and/or contact times in field trials with apparently reduced in situ efficacy. J . B . W R I G H T , I . RUSESKA A N D J . W . COSTERTON. 1991. In
INTRODUCTION
Legionella spp., which are the causative agents of Legionnaires’ disease and Pontiac fever (Glick et al. 1978), are also believed to occasionally play a contributory role in other respiratory diseases (Wright et al. 1988). Studies which have been carried out since the 1976 outbreak of Legionnaires’ disease in Philadelphia (McDade et al. 1977) have suggested that the primary mode of dispersion of the aetiological agent is an aerosol, inhalation of which may result in infection (Glick et al. 1978; Thacker et al. 1978; Mangione & Broome 1985). T h e putative source of the infecting bacteria has frequently been contaminated cooling tower water (Cordes et al. 1980; Dondero et al. 1980; Kaufmann et al. 1981) or contaminated potable water (States et al. 1987a; Vickers et al. 1987). Most of the individuals who contract a life-threatening Legionnaires’ disease are already debilitated in some way. Increased susceptibility may be the result of circumstances ranging from advancing age to immunosuppressive therapy (Sykes & Brazier 1988). Hospitals have therefore often been the sites of outbreaks of Legionnaires’ disease (e.g. Garbe et al. 1985; Timbury et al. 1986). This has led to the advancement of a number of protocols for controlling the
Correspondence t o : Dr J . B . Wright, Berz Paperchem, Inc., 7510 Baymeadows Wa,y, Jacksonville, FL 32256, USA.
population of legionellas in these and other large institutions (see below). Various protocols have been examined for the conttol of legionellas in potable water supplies, including chlorination (Swango et al. 1987; Helms et al. 1988), heat (Stout et al. 1986), and ultraviolet light (Muraca et al. 1987; Yamamoto et al. 1987). Attempts to control these bacteria in cooling towers is almost universally accomplished by the recommended, regular maintenance and cleaning of the towers in conjunction with regular biocide treatment (Wright et al. 1989a; Albrechtsen et al. 1990). A number of different biocides have been examined in vitro to determine their effectiveness against these bacteria (Grace et al. 1981; States et al. 1987b). Many biocides that showed promising results in the laboratory have been less efficacious in situ (e.g. by England et al. 1983; Fliermans & Harvey 1984). T h e present work was undertaken to determine reasons for the better survival of naturally occurring organisms vs those used in laboratory experiments. These experiments were carried out under conditions similar to those in which the biocides were designed to function, i.e. cooling tower conditions of high alkalinity and water hardness. T h e results indicated that one of the most important factors which need to be considered in studies of biocide efficacy is that surface-associated bacteria tend to be more resistant to biocides than the corresponding planktonic cells. Hence, the effectiveness of the biocide against these cells must also be examined.
532 J . B . WRIGHT ET AL
MATERIALS AND METHODS Bacteria
The organism used was a strain of Legionella pneumophila serogroup 1, monoclonal antibody type OXFORD 4032E. This strain was isolated by the Public Health Laboratory Service (PHLS), Porton Down, UK, from a cooling tower in Southampton, UK. The bacteria were typed by D r Jean Joly (Universiti: Laval, Canada) according to his published protocol (Joly 8z Winn 1984). The bacteria were stored on buffered charcoal yeast extract agar (BCYE) slopes at -70°C. The original strain was plated for purity and slopes were prepared from single colonies. T h e slopes were incubated at 37°C for 4 d to ensure heavy growth of the bacteria and then flash frozen in liquid nitrogen. Sufficient slopes were prepared and stored frozen until required, eliminating the need for numerous subcultures. Media
Buffered charcoal yeast extract agar (Gibco, Madison, WI) was used and was prepared according to the manufacturer's instructions. Liquid medium
The liquid medium employed for all experimental work was an artificially hardened water (AHW) containing (g/l in distilled water): MgCl, , 0.29; CaCI, . 2H20, 0.66; NaHCO, , 0.20; and supplemented with algal extract (Pope et al. 1982). The final p H of the medium was approximately 7.6 and it had a hardness of approximately 530 ppm as CaCO, . Water hardness was determined by the method described by Horwitz (1980). This formulation of artificially hardened water was chosen as it is typical of cooling tower waters in North America and Europe, according to industrial water treatment experts. Preparation of algal extract
The medium used for the cultivation of the blue-green algae used for the algal extract contained (g/l in distilled water): garden soil (200) and CaC03 (0.125). Fischerella spp. (Carolina Biological Supply, Burlington, NC) were then added to the steam-sterilized medium and allowed to grow under natural light at 25°C. After growth for several weeks, a volume of medium was removed and filtersterilized (0.45 p m ; MSI, Honeoye, NY). T h e filtrate constituted the algal extract, which was added to the AHW to a final concentration of 10%. Colonization apparatus
The device used to colonize the cooling tower component surface material (Douglas fir and polyvinyl chloride (PVC))
was the modified Robbins device (MRD), described in detail by Wright et al. (1989b). T h e device, essentially a tube with removable surfaces, consists of a perspex chamber (0.6 x 1.9 x 41.3 cm) with a number of removable sample ports along the top of the chamber. T h e ports have recessed bottom edges into which fit studs (with a face area of 0.5 cm2), made of various materials. T h e ports are designed such that the studs lie flush with the top of the flow chamber. Flow through the device can be halted for sampling and new, sterile studs can replace those removed for examination. T h e M R D was connected to a 2 1 reservoir by lengths of siliconized tubing (Tygon, Pittsburgh, PA). T h e bacterial culture was continuously recirculated through the M R D by a peristaltic cassette pump (Manostat; New York, NY), pumping at a rate of 240 ml/h. The pump could accommodate up to five devices operating simultaneously. To ensure adequate aeration of the closed system, each reservoir was connected to an Elite 800 fish tank pump (Rolf C. Hagen Corp., Mansfield, MA). Coionlzation procedure
The assembled MRDs and reservoirs were sterilized by ethylene oxide for 4.5 h. The apparatus was then allowed to stand overnight in a microbiological safety cabinet to allow the complete dissipation of any remaining ethylene oxide. Initially, the AHW was seeded with 10' L . pneumophila (measured with an O.D.,,, standard) and circulated through the device for 14 days. T o promote rapid colonization, replenish nutrients, and increase volume lost due to evaporation, the reservoirs received 5 x lo4 L. pneumophila in AHW every 5-6 d (twice during colonization). Sufficient medium (approximately 50-100 ml) was added to bring the total volume up to 900 ml. Sampling
T o sample the planktonic population 1 ml of the culture fluid was removed and serially diluted in sterile phosphate buffered saline (PBS; pH 7.0). A 0.1 ml volume of each dilution was plated, in duplicate, on BCYE plates. The plates were incubated at 37°C until colonies of L. pneumophila appeared, or for a maximum of 7 d. If colonies did not appear from the undiluted sample after incubation for 7 d, it was assumed that the number of viable bacteria in the sample was below detectable limits ( 10, cfu/ml). To sample the sessile population a sample port was removed aseptically and rinsed under a stream (10 ml) of sterile PBS to remove any loosely adherent bacteria. The sample surface was then scraped with a sterile scalpel blade and the scrapings, along with the sample, added to a tube containing 10 ml of sterile PBS. The stud could be ejected from the port by an alcohol-sterilized push rod. The tube and contents were ultrasonicated in a low output (50-60
BlOClDE SUSCEPTIBILITY OF L . PNEUMOPHlLA
Hz) bath ultrasonicator (Branson; Shelton, CT) for 5 min. This treatment was found to be effective in removing all bacteria from the surface of the stud and dispersing any aggregates of cells, without being detrimental to the viability of the recovered bacteria (Wright et al. 1989b). The contents of the tube were then serially diluted, plated, incubated and the colonies counted as described above. Tube contents were also filtered through a 0.45 pm acetate filter (MSI, Honeoye, NY) and the filter placed on BCYE agar. If no growth was seen on any of the plates after incubation for 7 d, the number of culturable cells was determined to be below detectable limits ( < 30/cm2). Bloclde testing
The biocides which were employed in these experiments were 2-bromo-4-nitropropane-1,3-diol(Bronopol [Baccillin LP], Bird Archer Inc., Coborg, ON) and a combination of 5-chloro-2-methyl-4-isothiazolin-3-one and 2-methyl-4isothiazolin-3-one (Kathon ; Rohm and Haas Company, Springhouse, PA). Bronopol was tested at final concentrations of 50 and 100 ppm, active ingredient; Kathon was tested at final concentrations of 9 and 30 ppm, active ingredient. In preparing the biocides for use, directions given by the respective firms were followed to ensure that final levels of active ingredient were accurate. In all the biocide efficacy experiments a biofilm was allowed to develop on the test surfaces for 14 d. Enough of the test biocides were then added to give the desired concentration of active ingredient in the 1 1 final volume. Before the biocide was added the planktonic and sessile populations were sampled. Similarly, samples were taken at 3, 6, 9, 12, 24 or 48 h after the addition of the biocide. The samples were diluted and plated immediately after sampling. Usually, both planktonic and sessile samples were taken at the indicated times. Experiments with each biocide were performed, with duplicate samplings, a minimum of five times to ensure accuracy and reduce error associated with variability in the extent of colonization of surfaces along the length of the MRD. It should be noted that the biocides were not inactivated prior to diluting or plating the samples so that the numbers of colonies counted represented the maximum number of culturable bacterial cells in the sample.
533
fixed in 5 % glutaraldehyde. After fixation, the surfaces were rinsed in cacodylate buffer and dehydrated, prior to being coated with gold and examined with a Hitachi S650 scanning electron microscope (Wright et al. 1989b). RESULTS
After two weeks of colonization in the MRD, the wood and PVC surfaces showed a substantial number of adherent L. pneumophila cells which remained at a constant level over the course of the experiment (Table 1). Figures 1 and 2 are scanning electron micrographs of the colonized fir and
Table 1 Sessile and planktonic population fluctuation in a modified Robbins device without the addition of biocide
Time (h)
Planktonic (cfu/mI)
Fir (cfu/cm2)
PVC (cfu/cm2)
0 6 24
7.51 f 0.18 7.57 f 0.16 7-65 f 0.16
6.49 f 0.18 6.26 & 0.24 6.29 & 0.30
5.46 f 0.21 5.44 k 0.19 5.45 f 0.34
All numbers are expressed as the log of the actual values.
Scannlng electron microscopy
The colonized materials were examined by scanning electron microscopy (SEM) to determine the extent of colonization and any effect which the biocides had on removing the biofilms from the surface of the stubs. Briefly, the surfaces were removed from the MRD, rinsed with PBS, and
Fig. 1 Scanning electron micrograph of Douglas fir after 14 days of colonization by Legionelfu pneumophifu. The arrows indicate microcolonies of the bacteria. Bar = 5 pm
534 J . B . WRIGHT ET A L .
Fig. 2 Scanning electron micrograph of polyvinyl chloride
following colonization by Legionella pneumophila. The arrow indicates a microcolony of cells. Bar = 5 p n
PVC, respectively, after 14 d colonization in the MRD. Adherent cells of L. pneumophila are visible in each micrograph. The arrows on the micrographs indicate microcolonies of bacteria which appear to be surrounded by an amorphous material. Previous studies have indicated that this material can be antibody stabilized with rabbit-raised antibodies against whole, formalin-killed L. pneumophila. Subsequently this material stained with ruthenium red
(Wright & Costerton 1986). These properties suggest that the material is an anionic polysaccharide (Luft 1971) material, termed ‘glycocalyx’ by Costerton et al. (1981). Table 2 shows the effect of adding 50 or 100 pprn of Bronopol to a pre-colonized system. T h e biocide was most active against the planktonic cells, with the 100 pprn dose reducing the number of viable bacteria by more than 4 logs after contact for 9 h. After 12 h of continuous exposure to 100 ppm Bronopol the number of viable bacteria was reduced to levels below detection ( < lo2 ml). T h e 50 pprn dosage required a much longer time to reduce the planktonic population to similar levels. Table 2 also shows that the biocide is much less effective against the adherent population. Longer contact times were required to have an effect on the sessile population, even with the 100 ppm dosage. However, after continuous exposure for 24 h, both the 50 and 100 ppm concentrations were successful in markedly reducing the number of culturable L. pneumophila on the surface of the fir disks (99.7 and 99.9%, respectively, based on cfu/cm2). After the same length of time the number of viable cells on the surface of the PVC disks was reduced to a level below detection ( < 30/cm2). Increasing the concentration of the biocide from 50 to 100 pprn only seemed to enhance the effect of the biocide on planktonic cells, as compared with adherent cells. Table 3 shows similar results for the effectiveness of an isothiazolin-containing biocide (Kathon). T h e biocide, at either 9 or 30 pprn active ingredient, was seen to have a more rapid effect against planktonic bacteria over a 24 h period than against the adherent population. Over this time, however, it was not found to be very effective against either the sessile or planktonic bacteria as significant numbers of culturable bacteria remained after the 24 h treatment. At 30 ppm, Kathon was no more effective against the L. pneumophila over the initial 24 h period than was the 9 pprn concentration. However, 30 ppm Kathon was successful at reducing the number of culturable cells
Table 2 Effect of 50 and 100 ppm Bronopol on planktonic and sessile populations of Legionella pneumophila
Planktonic
Fir
Time (h)
50 PPm
100 PPm
0 3 6 9 12 24
7.31 f 0.03 6.84 f 0.04 6.30 f 0.06 5.39 f 0.05 3.36 f 0.16 BDL
6.18 f 0.03 ND 4.32 f 0.04 1.70 f 0.07 BDL BDL
4.76 f 0.14 4.57 f 0.04 4.38 f 0.12 3.45 f 0.15 3.54 f 0.06 2.28 f 0.30
PVC
5.08 f 0.07 5.10 f 0.03 4.99 f 0.03 2.85 f 0-04 3.70 f 0.06 1.79 f 0.09
Planktonic population expressed as log cfu/ml; sessile population as log cfu/cm2. ND, Not determined; BDL, below detectable limits: planktonic < 102/ml; sessile < 30/cm2.
4.40 f 0.16 4.14 f 0.06 4.22 f 0.27 3.15 f 0.15 2.75 f 0.15 BDL
5.00 f 0.12 4.86 f 0.03 4.48 f 0.08 2.60 f 0.10 2.48 f 0.15 BDL
BlOClDE SUSCEPTIBILITY OF L . PNEUMOPHILA 535
Table 3 Effect of 9 and 30 ppm Kathon against sessile and planktonic populations of Legionella pneumophila
Planktonic Time (h)
9 PPm
PVC
Fir 30 PPm
30 PPm
9 PPm
30 PPm
9 PPm ~
0 3 6 9 24 48
7.37 f 0.04 6.15 f 0.01 5.85 0.01 5-35 k 0.07 3.53 k 0.02 2.24 f 0.20
7.62 f 0.01 ND
ND ND 3.46 _+ 0.01 BDL
5.70 _+ 0.04 5.04 f 0.01 4.51 f 0.02 4.47 _+ 0.06 4.44 0.03 3-24 f 0.12
6.09 f 0.03
ND ND ND 5.39 f 0.02 BDL
5.26 f 0.06 5.03 _+ 0.02 4.15 _+ 0.05 3.88 f 0.03 4.34 f 0.01 1.64 f 0.48
~
~~~~
5.68 0.02 ND ND ND 4.96 0.15 BDL
Planktonic results expressed as log cfu/ml; sessile results as log cfu/cm2. ND, Not determined; BDL, below detectable limits: planktonic < 102/ml;sessile < 30/cmZ.
(both planktonic and sessile) to levels below detection following continuous exposure to the biocide for 48 h.
DISCUSSION
Biocides have been used in cooling towers and other heatexchange systems for a number of years in an attempt to control the microorganisms which foul these systems. The fouling of such systems increases the cost of using them by decreasing their effectiveness at heat-exchange as well as increasing the frictional flow resistance. Because of the variety of fouling organisms, most commercial biocides have a broad spectrum of activity. After a link was esrablished between Legionnaires’ disease and the release of legionellas in aerosols from cooling waters, a substantial amount of work was done to examine the efficacy of these biocides against Legionella spp. This work led to the marketing of a number of compounds, including Bronopol and Kathon, which were purported to be effective against these pathogens, in addition to maintaining the general effectiveness of the biocides against other fouling organisms. The main thrust of the present work came from the knowledge of the resistance of other organisms to industrial biocides (Ruseska et al. 1982). Results from various investigators (Schofield & Locci 1985; Timbury et al. 1986; GIbourne & Dennis 1988; Wright et al. 1989b) have established that L. pneumophila grows in biofilms on the liquid-bathed surfaces of different types of water systems. These types of observations have led to recommendations about the importance of examining the effectiveness of biocides against biofilm-associated L. pneumophila (Wright et al. 1989a). Studies that used a simple system which allowed for the examination of L . pneumophila-colonized surfaces (in addition to aged cells, thought to be representative of biofilm cells (Page & Gaylarde 1990)) during the course of biocide treatment were a step towards a fuller understanding of the interaction of biofilms and biocides.
Initial experiments showed that it was necessary to recirculate the media through the MRD for 14 d to establish a reasonable number of adherent bacteria which had been sessile for long enough to equate them to established biofilm-associated organisms in field conditions. It was also necessary to add more organisms and nutrients to the system during this period to promote a satisfactory rate of colonization, since continuous culture conditions (similar to those of Keevil 1989) were not employed. Figures 1 and 2 show that in this system the organisms adhered to the fir and PVC disks and that they were frequently seen in microcolonies surrounded by an amorphous material, the glycocalyx, similar to that which we have reported in other laboratory systems as well as in operating water distribution systems (Wright et al. 1989b). In our observations of pure cultures of L. pneumophila growing on and adhering to surfaces, we have never noted the establishment of a confluent biofilm similar to those produced by other organisms. The lack of formation of confluent masses of adherent organisms may reflect the mode of growth of the organisms in ‘natural’ systems where they tend to grow in conjunction with other species of microorganisms, probably in multi-species consortia. However, since these organisms have also been described as existing in mono-specific microcolonies in situ (Wright et al. 1989b), we chose to examine the efficacy of biocides against pure cultures of L. pneumophila. T h e results reported here would be considered as the ‘best-case’ scenario for the ability of biocides to interfere with the growth of these organisms. If these organisms existed within thicker biofilms the biocides would be suspected of being less effective against the organisms than if they were in smaller microcolonies (see Keevil et al. 1990). In Canada, Bronopol is recommended for commercial use at a concentration of 50 ppm, or at 100 pprn for more heavily fouled systems. Kathon was used at concentrations of 9 and 30 ppm. The 9 ppm dosage was suggested as being effective in an earlier publication (McCoy et al. 1986) and
536 J . B . W R I G H T ET A L .
the 30 ppm concentration is recommended by the manufacturers for use in fouled systems. The results of the biocide challenge (Tables 2 and 3) show that both biocides were quite effective at sharply reducing, and possibly eliminating, viable L. pneumophila cells from the bulk fluid (planktonic population). Bronopol appeared to be more effective than Kathon, at the concentrations recommended by the manufacturers. Since the effective concentration of biocide within a cooling tower falls off rapidly as a result of loss through drift and dilution through the addition of make-up water, the more rapidly a biocide works the more effective it is likely to be. Both biocides were also effective in reducing the number of viable, adherent bacteria. While the results indicated that the biocides were capable of reducing the numbers of sessile bacteria by 8&99% (based on viable counts) within a 9-24 h period, this still left a relatively large population of viable bacteria on the surfaces. T h e presence of this population would serve as an excellent reservoir of bacteria to recontaminate the system after cessation of the biocide dosing. Additionally, the biocide challenges were seen to have little effect on removing the dead bacterial cells from the surfaces as assessed by electron microscopy (micrographs not shown). The presence of bacterial cells or cell products has been shown to promote the more rapid and tenacious recolonization of surfaces in other systems (Nickels et al. 1981). In field studies where the efficacy of various biocides has been tested for their ability to deplete and control L. pneumophila populations, the surfaceassociated mode of growth of L. pneumophila may be one of the factors involved in the poor performance of the biocide as compared with the results expected from laboratory tests where the adherent population of bacteria was not examined. At least two major hypotheses have been published which help explain the generally noted differences in the antimicrobial susceptibilities between adherent and planktonic bacteria. Costerton (1984) proposed that the resistance to antimicrobial killing of sessile bacteria (vs their planktonic counterparts) arises by virtue of their ability to grow adherent to surfaces in glycocalyx-enclosed microcolonies. Brown et al. (1988) also proposed that the increased resistance of sessile bacteria to antibiotics is a result of their slower growth rate. In this particular case, neither of these hypotheses appear able to explain the results on their own. The situation is probably more complex as a result of the bacteria growing in glycocalyxenclosed microcolonies as well as both populations of bacteria growing at a very slow rate as shown by the need for replenishing the system to promote colonization and growth. In a practical sense, the results indicate that the sessile population must be taken into account when evaluating the
eficacy of biocides against cooling tower Legionella spp. At the same time particular attention must be paid to the test conditions so that the experimental system reflects the conditions under which the biocide will be functioning. More experimental work will be required to elucidate the mechanism by which sessile bacteria are afforded greater resistance to the decontamination procedures which are effective against the planktonic bacteria. These types of studies will be best addressed utilizing the carefully controlled environments described by Keevil (1989) for the growth of biofilms. ACKNOWLEDGEMENTS T h e authors wish to thank Glynnis Broadbent and Urvashi Sabharwal for their expert assistance with photography and electron microscopy, and to Bird Archer Co. and Rohm and Haas Co. for the supply of biocides and technical information. This work was supported by grants from the Alberta Heritage Foundation for Medical Research. REFERENCES A L B R E C H T S E N , O., BARTLETT, C.L.R., B E T T E L H E I M , K.A., B O R N S T E I N., N , C O L B O U R NJE. S, . , F L E U R E T T E ,
J . , F R A S E R ,D.W., J O L Y , J.R., K A L L I N G S I, . , M A R T I N - B O U R G NC., O N ,PROSOWSKY, S.B., R O W B O THAM, T.J., STOLWIJK J.A.S., , TARTAKOVSK I .Y S ., , & Y A E U U C H I , E. (1990) Epidemiology, prevention and control of legionellosis : Memorandum from a WHO meeting. Bulletin of the World Health Organisation 68, 155-164.
B R O W NM.R.W., , ALLISON,D.A. & G I L B E R TP. , (1988) Resistance of bacterial biofilms to antibiotics: A growth-rate related effect ? Journal of Antimicrobial Chemotherapy 22, 777778.
C O L B O U R NJE. S, . & D E N N I S ,P.J. (1988) Legionella: A biofilm organism in engineered water systems? In: Biodeterioration 7: Seventh International Biodeterioration Symposium. ed. Houghton, D.R. et al. pp. 3 H 2 . Barking: Elsevier Applied Science. CORDES,L.G., F R A S E RD.W., , S K A L I YP., , PERLINO, C.A., ELSEA,W.R., M A L L I S O NG,. F . & H A Y E SP.S. , (1980) Legionnaires’ disease outbreak at an Atlanta, Georgia country club: Evidence for spread from an evaporative condenser. American Journal of Epidemiology 111,425-53 1. COSTERTON, J.W., I R V I NR , . T . & C H E N G K.-J. , (1981) The bacterial glycocalyx in nature and disease. Annual Reviews of Microbiology 35, 299-324.
COSTERTON, J.W. (1984) The formation of biocide-resistant biofilms in industrial, natural and medical systems. Developments in Industrial Microbiology 25, 363-372.
D O N D E R OT, .J . , R E N D T R O FR.C., F, M A L L I S O NG, . F . , WEEKS,R . M . , L E V Y J, . S . , WONG,E.W. & S C H A F F N E R , W. (1980) An outbreak of Legionnaires’ disease associated with a contaminated air-conditioning cooling tower. New England Journal of Medicine 302, 365-370.
BlOClDE SUSCEPTIBILITY OF L. PNEUMOPHILA 537
E N G L A N 111, D A.C., F R A S E R D ,. W . , M A L L I S O NG, . F . , M A C K E LD.C., , S K A L I YP. , & G O R M A NG, . W . (1983) Failure of Legionella pneumophila sensitivities to predict culture results from disinfectant-treated air-conditioning cooling towers. Applied and Environmental Microbiology 43,240-244. F L I E R M A N C.B. S, & HARVEY R,. S . (1984) Effectiveness of I-bromo-3-chlor0-5,Sdimethylhydantoin against Legionella pneumophila in a cooling tower. Applied and Environmental Microbiology 47, 1307-1310. G A R B E ,P.L., D A V I S ,B.J., WEISFELD,J . S . , M A R K O W I T Z , L., M I N E R ,P . , G A R R I T YF, . , B A R B A R E JE.,M . & REINGOLD, A.L. (1985) Nosocomial Legionnaires’ disease : Epidemiologic demonstration of cooling towers as a source. Journal of the American Medical Association 254, 521-
524. G L I C K ,T . H . , G R E G G ,M.B., B E R M A NB., , MALLISON, G., RHODES,W . W . & KASSANOFF, L . (1978) Pontiac fever. An epidemic of unknown etiology in a health department. I. Clinical and epidemiological aspects. American Journal of Epidemiology 107, 14%160. G R A C ER.D., , D E W A RN.E., , B A R N E SW , . G . & HODGES, G. R. (1981) Susceptibility of Legionella pneumophila to three cooling tower microbiocides. Applied and Environmental Microbiology 41, 233-236. HELMS, C . M . , M A S S A N A R IR, . M . , W E N Z E L , R.P., PFELLER,M.A., M O Y E R , N . P . , H U L L , N . & THE L E GI o N E L L A M O NI T O R I N G C o M M I T T E E ( 1988) Legionnaires’ disease associated with a hospital water system: A fiveyear progress report on continuous hyperchlorination. Journal of the American Medical Association 259, 2423-2427. HORWITZ, W. ( e d . ) (1980) Oficial Methods of Analysis of the Association of Oflcial Analytical Chemists. 4.027-4.028. Washington, DC : Association of Official Analytical Chemists. J O L Y , J . R . & W I N N ,W . C . J R (1984)Correlation of subtypes of Legionella pneumophila defined by monoclonal antibodies with epidemiological classification of cases and environmental sources. Journal of Infectious Diseases 150,667-67 1. K A U F M A N NA , . F . , M C D A D E , J . E . , P A T T O N , C.M., B E N N E T TJ.V., , S K A L I Y P, . , F E E L E Y J, . C . , A N D E R SON,D . C . , P O T T E RM , . E . , N E W H O U S EV, . F . , G R E G G , M.B. & B R A C H M A PN.,S . (1981)Pontiac fever: Isolation of the etiologic agent (Legionella pneumophila) and demonstration of its mode of transmission. American Journa! of Epidemiomology 114, 337-347. K E E V I L ,C . W. (1989) Continuous culture studies of Legionella pneumophila. Journal of Applied Bacteriology 67, x. K E E V I LC.W., , M A C K E R N E SC.W. S, & C O L B O U R N JE.,S . (1990) Biocide treatment of biofilms. International Biodeterioration 26, 169-179. L U F T , J . H . (1971) Ruthenium red and ruthenium violet. I. Chemistry, purification, methods of use for electron microscopy and mechanism of action. Anatomical Record 171, 347-368. M A N G I O N EE.J. , & BROOME,C . V . (1985) Legionellosis: Epidemiology. In Legionellosis ed. Katz, S.M. pp. 6tL73. Boca Raton: Chemical Rubber Company Press. M c C o y , W . F . , W I R E M A NJ ., W . & L A S H E NE, . S . (1986) Efficacy of methychloro/methylisothiazolone biocide against
Legionella pneumophila in cooling tower water. Journal of Industrial Microbiology 1, 4%56. M C D A D E J, . E . , S H E P A R DC, . C . , F R A S E R D,. W . , TSAI, T . R . , REDUS,M . A . , DOWDLE, W.R. & T H EL A B O R ATORY INVESTIGATIO TNE A M (1977)Legionnaires’ disease: Isolation of a bacterium and demonstration of its role in other respiratory disease. New England Journal of Medicine 297,
1 197- 1203. M U R A C AP., , S T O U T J.W. , & Y u , V . L . (1987)Comparative assessment of chlorine, heat, ozone, and UV light for killing Legionella pneumophila within a model plumbing system. Applied and Environmental Microbiology 53,447453. N I C K E L S ,J . S . , BOBBIE, R . J . , L O T T , D.F., M A R T Z , R . F . , BENSON,P.H. & W H I T E , D . C . (1981) Effect of manual brush cleaning on biomass and community structure of microfouling film formed on aluminum and titanium surfaces exposed to rapidly flowing seawater. Applied and Environmental Microbiology 41, 1442-1453. P A G E ,S . & G A Y L A R D EC., (1990) Biocide activity against Legionella and Pseudomonas. International Biodeterioration 26,
139-148. POPE, D . H . , S O R A C C OR.J., , GILL, H.K. & FLIERM A N S , C.B. (1982)Growth of Legionella pneumophila in twomembered cultures with green algae and cyanobacteria. Current Microbiology 7 , 319-322. RUSESKA, I., R O B B I N SJ,. , L A S H E NE ,. S . & COSTERTON, J .W . (1982) Biocide testing against corrosion-causing oilfield bacteria helps control plugging. Oil and Gas Journal 8, 253-
264. S C H O F I E LGD. M , . & L O C C I ,R. (1985)Colonization of components of a model hot water system by Legionella pneumophila. Journal of Applied Bacteriology 58, 151-162. S T A T E S ,S.J., C O N L E Y ,L . F . , K U C H T A J, . M . , O L E C K , B.M., L I P O V I C HM.J., , WOLFORD,R.S., W A D O W S K Y , R.M., MCNAMARA A.M., , S Y K O RJA. L, . , K E L E T I ,G . & Y E E , R. B . (1987a)Survival and multiplication of Legionella pneumophila in municipal drinking water systems. Applied and Environmental Microbiology 53, 979-986. STATES, S . J . , C O N L E Y , L . F . , T O W N E R , S . G . , WOLFORD, S . , STEPHENSON, T.E., MCNAMARA, A.M., W A D O W S K Y R,. M . & YEE, R.B. (1987b) An alkaline approach to treating cooling towers for control of Legionella pneumophila. Applied and Environmental Microbiology 53, 1775-
1779. S T O U TJ, . E . , BEST, M . G . & Yu, V . L . (1986) Susceptibility of members of the family Legionellaceae to thermal stress : Implications for heat eradication methods in water distribution systems. Applied and Environmental Microbiology 52, 396-399. S W A N G OL.J., , W I L T , G . R . , K I L L E NA , .D., WILLIAMS, D . E . & W O R L E Y ,S . D . (1987) Inactivation of Legionella pneumophila by hypochlorite and an organic chloramine. Applied and Environmental Microbiology 53, 2983-2986. S Y K E S ,J . M . & B R A Z I E R A , . M . (1988) Assessment and control of risks from Legionnaires’ disease. Annals of Occupational Hygiene 32, 63-67. T H A C K ESR . B, . , B E N N E T TJ, . V . , T S A I , T . F . , F R A S E R , D . W . , M C D A D E ,J . E . , S H E P A R DC , .C., WILLIAMS,
538 J . B . W R I G H T E r A L .
K . H . J R , S T U A R TW , .H., DULL, H.B. & EICKHOFF, T. C . (1978) An outbreak in 1965 of severe respiratory illness caused by the Legionnaires’ disease bacterium. Journal of Infectious Diseases 138, 512-519. TIMBURY M,. C . , D O N A L D S O NJ, . R . , M C C A R T N E Y , A.C., F A L L O NR.J., , S L E I G H , J . D . , L Y O N , D., O R A N G EG.V., , B A I R D D.R., , W I N T E R J, . & W I L S O N , T . S . (1986) Oubreak of Legionnaires’ disease in Glasgow Royal Infirmary : Microbiological aspects. Journal of Hygiene 97, 393403. V I C K E R SR.M., , Yu, V.L., H A N N A ,S., M U R A C AP, . , D I V E N W., , C A R M E NN, . & T A Y L OFR. B , . (1987) Determinants of Legionella pneumophila contamination of water distribution systems : 15-hospital prospective study. Infection Control, 8, 357-363. W R I G H TJ.B. , & COSTERTON J .,W . (1986) Visualization of the glycocalyx of Legionella pneumophila. Abstracts of the Annual Meeting of the Alberta Heritage Foundation f o r Medical Research p. 26. Edmonton: Alberta Heritage Foundation for Medical Research.
W R I G H T ,J . B . , A T H A R , M . A . , V A N O L M ,T . M . , WOOTL I F F , J . S . 8~ C O S T E R T O N J ., W . (1988) Atypical legionellosis: Isolation of Legionella pneumophila serogroup 1 from a patient with aspiration pneumonia. Journal of Hospital Infection 13, 187-190. W R I G H T A.E., , B A R T L E T TC, . L . R . , B R U N D R E TG T ,. W . , H I L L , E.C., H U G O , W.B., H U T C H I S O N J, . G . P . , J O N E S ,W . P . , L E E , J., P I K E , E . B . , S P O O N E RD, . F . , TILLETT H ,. E . & P E N N ,R . G . (1989a) Report ofthe Expert Advisory Committee on Biocides. Department of Health, London : HMSO. W R I G H T ,J . B . , RUSESKA,I . , A T H A R , M.A., C O R B E T T , S. & COSTERTON,J.W. (1989b) Legionella pneumophila grows adherent to surfaces in vitro and in situ. Infection Control and Hospital Epidemiology 10, 4 0 8 4 1 5 . YAMAMOTO H,. , U R A K A M I ., , N A K O N OK , . , I K E D O ,M . & Y A B U U C H IE, . (1987) Effect of FLONLIZER ultraviolet sterilizer on Legionella species inhabiting cooling tower water. Microbiology and Immunology 31,745-752.
Decreased biocide susceptibility of adherent Legionella pneumophila J.B. Wright, 1. Ruseska and J.W. Costerton Department of Biological Sciences, University of Calgary, CalgaryvAlberta, Canada 3513/10/90:accepted 1 June 1991
a study of the in vitro effectiveness of biocides against Legionella pneumophila, some aspects of the cooling tower environment were replicated in the laboratory, paying particular attention to water hardness and pH. Pieces of Douglas fir and polyvinyl chloride were colonized in a recirculating system and the comparative efficacy of two biocides (Bronopol and Kathon) against the sessile and planktonic populations was examined. While the biocides were relatively effective against the planktonic L . pneumophila population over a short period of time (minimum 9-12 h), substantially longer periods of time (maximum > 48 h) were required to reduce the number of cultivable bacteria to below detectable levels in the adherent population. The results indicate that failure to monitor the sessile population of L. pneumophila in laboratory studies of biocides may result in the use of incorrect dosages and/or contact times in field trials with apparently reduced in situ efficacy. J . B . W R I G H T , I . RUSESKA A N D J . W . COSTERTON. 1991. In
INTRODUCTION
Legionella spp., which are the causative agents of Legionnaires’ disease and Pontiac fever (Glick et al. 1978), are also believed to occasionally play a contributory role in other respiratory diseases (Wright et al. 1988). Studies which have been carried out since the 1976 outbreak of Legionnaires’ disease in Philadelphia (McDade et al. 1977) have suggested that the primary mode of dispersion of the aetiological agent is an aerosol, inhalation of which may result in infection (Glick et al. 1978; Thacker et al. 1978; Mangione & Broome 1985). T h e putative source of the infecting bacteria has frequently been contaminated cooling tower water (Cordes et al. 1980; Dondero et al. 1980; Kaufmann et al. 1981) or contaminated potable water (States et al. 1987a; Vickers et al. 1987). Most of the individuals who contract a life-threatening Legionnaires’ disease are already debilitated in some way. Increased susceptibility may be the result of circumstances ranging from advancing age to immunosuppressive therapy (Sykes & Brazier 1988). Hospitals have therefore often been the sites of outbreaks of Legionnaires’ disease (e.g. Garbe et al. 1985; Timbury et al. 1986). This has led to the advancement of a number of protocols for controlling the
Correspondence t o : Dr J . B . Wright, Berz Paperchem, Inc., 7510 Baymeadows Wa,y, Jacksonville, FL 32256, USA.
population of legionellas in these and other large institutions (see below). Various protocols have been examined for the conttol of legionellas in potable water supplies, including chlorination (Swango et al. 1987; Helms et al. 1988), heat (Stout et al. 1986), and ultraviolet light (Muraca et al. 1987; Yamamoto et al. 1987). Attempts to control these bacteria in cooling towers is almost universally accomplished by the recommended, regular maintenance and cleaning of the towers in conjunction with regular biocide treatment (Wright et al. 1989a; Albrechtsen et al. 1990). A number of different biocides have been examined in vitro to determine their effectiveness against these bacteria (Grace et al. 1981; States et al. 1987b). Many biocides that showed promising results in the laboratory have been less efficacious in situ (e.g. by England et al. 1983; Fliermans & Harvey 1984). T h e present work was undertaken to determine reasons for the better survival of naturally occurring organisms vs those used in laboratory experiments. These experiments were carried out under conditions similar to those in which the biocides were designed to function, i.e. cooling tower conditions of high alkalinity and water hardness. T h e results indicated that one of the most important factors which need to be considered in studies of biocide efficacy is that surface-associated bacteria tend to be more resistant to biocides than the corresponding planktonic cells. Hence, the effectiveness of the biocide against these cells must also be examined.
532 J . B . WRIGHT ET AL
MATERIALS AND METHODS Bacteria
The organism used was a strain of Legionella pneumophila serogroup 1, monoclonal antibody type OXFORD 4032E. This strain was isolated by the Public Health Laboratory Service (PHLS), Porton Down, UK, from a cooling tower in Southampton, UK. The bacteria were typed by D r Jean Joly (Universiti: Laval, Canada) according to his published protocol (Joly 8z Winn 1984). The bacteria were stored on buffered charcoal yeast extract agar (BCYE) slopes at -70°C. The original strain was plated for purity and slopes were prepared from single colonies. T h e slopes were incubated at 37°C for 4 d to ensure heavy growth of the bacteria and then flash frozen in liquid nitrogen. Sufficient slopes were prepared and stored frozen until required, eliminating the need for numerous subcultures. Media
Buffered charcoal yeast extract agar (Gibco, Madison, WI) was used and was prepared according to the manufacturer's instructions. Liquid medium
The liquid medium employed for all experimental work was an artificially hardened water (AHW) containing (g/l in distilled water): MgCl, , 0.29; CaCI, . 2H20, 0.66; NaHCO, , 0.20; and supplemented with algal extract (Pope et al. 1982). The final p H of the medium was approximately 7.6 and it had a hardness of approximately 530 ppm as CaCO, . Water hardness was determined by the method described by Horwitz (1980). This formulation of artificially hardened water was chosen as it is typical of cooling tower waters in North America and Europe, according to industrial water treatment experts. Preparation of algal extract
The medium used for the cultivation of the blue-green algae used for the algal extract contained (g/l in distilled water): garden soil (200) and CaC03 (0.125). Fischerella spp. (Carolina Biological Supply, Burlington, NC) were then added to the steam-sterilized medium and allowed to grow under natural light at 25°C. After growth for several weeks, a volume of medium was removed and filtersterilized (0.45 p m ; MSI, Honeoye, NY). T h e filtrate constituted the algal extract, which was added to the AHW to a final concentration of 10%. Colonization apparatus
The device used to colonize the cooling tower component surface material (Douglas fir and polyvinyl chloride (PVC))
was the modified Robbins device (MRD), described in detail by Wright et al. (1989b). T h e device, essentially a tube with removable surfaces, consists of a perspex chamber (0.6 x 1.9 x 41.3 cm) with a number of removable sample ports along the top of the chamber. T h e ports have recessed bottom edges into which fit studs (with a face area of 0.5 cm2), made of various materials. T h e ports are designed such that the studs lie flush with the top of the flow chamber. Flow through the device can be halted for sampling and new, sterile studs can replace those removed for examination. T h e M R D was connected to a 2 1 reservoir by lengths of siliconized tubing (Tygon, Pittsburgh, PA). T h e bacterial culture was continuously recirculated through the M R D by a peristaltic cassette pump (Manostat; New York, NY), pumping at a rate of 240 ml/h. The pump could accommodate up to five devices operating simultaneously. To ensure adequate aeration of the closed system, each reservoir was connected to an Elite 800 fish tank pump (Rolf C. Hagen Corp., Mansfield, MA). Coionlzation procedure
The assembled MRDs and reservoirs were sterilized by ethylene oxide for 4.5 h. The apparatus was then allowed to stand overnight in a microbiological safety cabinet to allow the complete dissipation of any remaining ethylene oxide. Initially, the AHW was seeded with 10' L . pneumophila (measured with an O.D.,,, standard) and circulated through the device for 14 days. T o promote rapid colonization, replenish nutrients, and increase volume lost due to evaporation, the reservoirs received 5 x lo4 L. pneumophila in AHW every 5-6 d (twice during colonization). Sufficient medium (approximately 50-100 ml) was added to bring the total volume up to 900 ml. Sampling
T o sample the planktonic population 1 ml of the culture fluid was removed and serially diluted in sterile phosphate buffered saline (PBS; pH 7.0). A 0.1 ml volume of each dilution was plated, in duplicate, on BCYE plates. The plates were incubated at 37°C until colonies of L. pneumophila appeared, or for a maximum of 7 d. If colonies did not appear from the undiluted sample after incubation for 7 d, it was assumed that the number of viable bacteria in the sample was below detectable limits ( 10, cfu/ml). To sample the sessile population a sample port was removed aseptically and rinsed under a stream (10 ml) of sterile PBS to remove any loosely adherent bacteria. The sample surface was then scraped with a sterile scalpel blade and the scrapings, along with the sample, added to a tube containing 10 ml of sterile PBS. The stud could be ejected from the port by an alcohol-sterilized push rod. The tube and contents were ultrasonicated in a low output (50-60
BlOClDE SUSCEPTIBILITY OF L . PNEUMOPHlLA
Hz) bath ultrasonicator (Branson; Shelton, CT) for 5 min. This treatment was found to be effective in removing all bacteria from the surface of the stud and dispersing any aggregates of cells, without being detrimental to the viability of the recovered bacteria (Wright et al. 1989b). The contents of the tube were then serially diluted, plated, incubated and the colonies counted as described above. Tube contents were also filtered through a 0.45 pm acetate filter (MSI, Honeoye, NY) and the filter placed on BCYE agar. If no growth was seen on any of the plates after incubation for 7 d, the number of culturable cells was determined to be below detectable limits ( < 30/cm2). Bloclde testing
The biocides which were employed in these experiments were 2-bromo-4-nitropropane-1,3-diol(Bronopol [Baccillin LP], Bird Archer Inc., Coborg, ON) and a combination of 5-chloro-2-methyl-4-isothiazolin-3-one and 2-methyl-4isothiazolin-3-one (Kathon ; Rohm and Haas Company, Springhouse, PA). Bronopol was tested at final concentrations of 50 and 100 ppm, active ingredient; Kathon was tested at final concentrations of 9 and 30 ppm, active ingredient. In preparing the biocides for use, directions given by the respective firms were followed to ensure that final levels of active ingredient were accurate. In all the biocide efficacy experiments a biofilm was allowed to develop on the test surfaces for 14 d. Enough of the test biocides were then added to give the desired concentration of active ingredient in the 1 1 final volume. Before the biocide was added the planktonic and sessile populations were sampled. Similarly, samples were taken at 3, 6, 9, 12, 24 or 48 h after the addition of the biocide. The samples were diluted and plated immediately after sampling. Usually, both planktonic and sessile samples were taken at the indicated times. Experiments with each biocide were performed, with duplicate samplings, a minimum of five times to ensure accuracy and reduce error associated with variability in the extent of colonization of surfaces along the length of the MRD. It should be noted that the biocides were not inactivated prior to diluting or plating the samples so that the numbers of colonies counted represented the maximum number of culturable bacterial cells in the sample.
533
fixed in 5 % glutaraldehyde. After fixation, the surfaces were rinsed in cacodylate buffer and dehydrated, prior to being coated with gold and examined with a Hitachi S650 scanning electron microscope (Wright et al. 1989b). RESULTS
After two weeks of colonization in the MRD, the wood and PVC surfaces showed a substantial number of adherent L. pneumophila cells which remained at a constant level over the course of the experiment (Table 1). Figures 1 and 2 are scanning electron micrographs of the colonized fir and
Table 1 Sessile and planktonic population fluctuation in a modified Robbins device without the addition of biocide
Time (h)
Planktonic (cfu/mI)
Fir (cfu/cm2)
PVC (cfu/cm2)
0 6 24
7.51 f 0.18 7.57 f 0.16 7-65 f 0.16
6.49 f 0.18 6.26 & 0.24 6.29 & 0.30
5.46 f 0.21 5.44 k 0.19 5.45 f 0.34
All numbers are expressed as the log of the actual values.
Scannlng electron microscopy
The colonized materials were examined by scanning electron microscopy (SEM) to determine the extent of colonization and any effect which the biocides had on removing the biofilms from the surface of the stubs. Briefly, the surfaces were removed from the MRD, rinsed with PBS, and
Fig. 1 Scanning electron micrograph of Douglas fir after 14 days of colonization by Legionelfu pneumophifu. The arrows indicate microcolonies of the bacteria. Bar = 5 pm
534 J . B . WRIGHT ET A L .
Fig. 2 Scanning electron micrograph of polyvinyl chloride
following colonization by Legionella pneumophila. The arrow indicates a microcolony of cells. Bar = 5 p n
PVC, respectively, after 14 d colonization in the MRD. Adherent cells of L. pneumophila are visible in each micrograph. The arrows on the micrographs indicate microcolonies of bacteria which appear to be surrounded by an amorphous material. Previous studies have indicated that this material can be antibody stabilized with rabbit-raised antibodies against whole, formalin-killed L. pneumophila. Subsequently this material stained with ruthenium red
(Wright & Costerton 1986). These properties suggest that the material is an anionic polysaccharide (Luft 1971) material, termed ‘glycocalyx’ by Costerton et al. (1981). Table 2 shows the effect of adding 50 or 100 pprn of Bronopol to a pre-colonized system. T h e biocide was most active against the planktonic cells, with the 100 pprn dose reducing the number of viable bacteria by more than 4 logs after contact for 9 h. After 12 h of continuous exposure to 100 ppm Bronopol the number of viable bacteria was reduced to levels below detection ( < lo2 ml). T h e 50 pprn dosage required a much longer time to reduce the planktonic population to similar levels. Table 2 also shows that the biocide is much less effective against the adherent population. Longer contact times were required to have an effect on the sessile population, even with the 100 ppm dosage. However, after continuous exposure for 24 h, both the 50 and 100 ppm concentrations were successful in markedly reducing the number of culturable L. pneumophila on the surface of the fir disks (99.7 and 99.9%, respectively, based on cfu/cm2). After the same length of time the number of viable cells on the surface of the PVC disks was reduced to a level below detection ( < 30/cm2). Increasing the concentration of the biocide from 50 to 100 pprn only seemed to enhance the effect of the biocide on planktonic cells, as compared with adherent cells. Table 3 shows similar results for the effectiveness of an isothiazolin-containing biocide (Kathon). T h e biocide, at either 9 or 30 pprn active ingredient, was seen to have a more rapid effect against planktonic bacteria over a 24 h period than against the adherent population. Over this time, however, it was not found to be very effective against either the sessile or planktonic bacteria as significant numbers of culturable bacteria remained after the 24 h treatment. At 30 ppm, Kathon was no more effective against the L. pneumophila over the initial 24 h period than was the 9 pprn concentration. However, 30 ppm Kathon was successful at reducing the number of culturable cells
Table 2 Effect of 50 and 100 ppm Bronopol on planktonic and sessile populations of Legionella pneumophila
Planktonic
Fir
Time (h)
50 PPm
100 PPm
0 3 6 9 12 24
7.31 f 0.03 6.84 f 0.04 6.30 f 0.06 5.39 f 0.05 3.36 f 0.16 BDL
6.18 f 0.03 ND 4.32 f 0.04 1.70 f 0.07 BDL BDL
4.76 f 0.14 4.57 f 0.04 4.38 f 0.12 3.45 f 0.15 3.54 f 0.06 2.28 f 0.30
PVC
5.08 f 0.07 5.10 f 0.03 4.99 f 0.03 2.85 f 0-04 3.70 f 0.06 1.79 f 0.09
Planktonic population expressed as log cfu/ml; sessile population as log cfu/cm2. ND, Not determined; BDL, below detectable limits: planktonic < 102/ml; sessile < 30/cm2.
4.40 f 0.16 4.14 f 0.06 4.22 f 0.27 3.15 f 0.15 2.75 f 0.15 BDL
5.00 f 0.12 4.86 f 0.03 4.48 f 0.08 2.60 f 0.10 2.48 f 0.15 BDL
BlOClDE SUSCEPTIBILITY OF L . PNEUMOPHILA 535
Table 3 Effect of 9 and 30 ppm Kathon against sessile and planktonic populations of Legionella pneumophila
Planktonic Time (h)
9 PPm
PVC
Fir 30 PPm
30 PPm
9 PPm
30 PPm
9 PPm ~
0 3 6 9 24 48
7.37 f 0.04 6.15 f 0.01 5.85 0.01 5-35 k 0.07 3.53 k 0.02 2.24 f 0.20
7.62 f 0.01 ND
ND ND 3.46 _+ 0.01 BDL
5.70 _+ 0.04 5.04 f 0.01 4.51 f 0.02 4.47 _+ 0.06 4.44 0.03 3-24 f 0.12
6.09 f 0.03
ND ND ND 5.39 f 0.02 BDL
5.26 f 0.06 5.03 _+ 0.02 4.15 _+ 0.05 3.88 f 0.03 4.34 f 0.01 1.64 f 0.48
~
~~~~
5.68 0.02 ND ND ND 4.96 0.15 BDL
Planktonic results expressed as log cfu/ml; sessile results as log cfu/cm2. ND, Not determined; BDL, below detectable limits: planktonic < 102/ml;sessile < 30/cmZ.
(both planktonic and sessile) to levels below detection following continuous exposure to the biocide for 48 h.
DISCUSSION
Biocides have been used in cooling towers and other heatexchange systems for a number of years in an attempt to control the microorganisms which foul these systems. The fouling of such systems increases the cost of using them by decreasing their effectiveness at heat-exchange as well as increasing the frictional flow resistance. Because of the variety of fouling organisms, most commercial biocides have a broad spectrum of activity. After a link was esrablished between Legionnaires’ disease and the release of legionellas in aerosols from cooling waters, a substantial amount of work was done to examine the efficacy of these biocides against Legionella spp. This work led to the marketing of a number of compounds, including Bronopol and Kathon, which were purported to be effective against these pathogens, in addition to maintaining the general effectiveness of the biocides against other fouling organisms. The main thrust of the present work came from the knowledge of the resistance of other organisms to industrial biocides (Ruseska et al. 1982). Results from various investigators (Schofield & Locci 1985; Timbury et al. 1986; GIbourne & Dennis 1988; Wright et al. 1989b) have established that L. pneumophila grows in biofilms on the liquid-bathed surfaces of different types of water systems. These types of observations have led to recommendations about the importance of examining the effectiveness of biocides against biofilm-associated L. pneumophila (Wright et al. 1989a). Studies that used a simple system which allowed for the examination of L . pneumophila-colonized surfaces (in addition to aged cells, thought to be representative of biofilm cells (Page & Gaylarde 1990)) during the course of biocide treatment were a step towards a fuller understanding of the interaction of biofilms and biocides.
Initial experiments showed that it was necessary to recirculate the media through the MRD for 14 d to establish a reasonable number of adherent bacteria which had been sessile for long enough to equate them to established biofilm-associated organisms in field conditions. It was also necessary to add more organisms and nutrients to the system during this period to promote a satisfactory rate of colonization, since continuous culture conditions (similar to those of Keevil 1989) were not employed. Figures 1 and 2 show that in this system the organisms adhered to the fir and PVC disks and that they were frequently seen in microcolonies surrounded by an amorphous material, the glycocalyx, similar to that which we have reported in other laboratory systems as well as in operating water distribution systems (Wright et al. 1989b). In our observations of pure cultures of L. pneumophila growing on and adhering to surfaces, we have never noted the establishment of a confluent biofilm similar to those produced by other organisms. The lack of formation of confluent masses of adherent organisms may reflect the mode of growth of the organisms in ‘natural’ systems where they tend to grow in conjunction with other species of microorganisms, probably in multi-species consortia. However, since these organisms have also been described as existing in mono-specific microcolonies in situ (Wright et al. 1989b), we chose to examine the efficacy of biocides against pure cultures of L. pneumophila. T h e results reported here would be considered as the ‘best-case’ scenario for the ability of biocides to interfere with the growth of these organisms. If these organisms existed within thicker biofilms the biocides would be suspected of being less effective against the organisms than if they were in smaller microcolonies (see Keevil et al. 1990). In Canada, Bronopol is recommended for commercial use at a concentration of 50 ppm, or at 100 pprn for more heavily fouled systems. Kathon was used at concentrations of 9 and 30 ppm. The 9 ppm dosage was suggested as being effective in an earlier publication (McCoy et al. 1986) and
536 J . B . W R I G H T ET A L .
the 30 ppm concentration is recommended by the manufacturers for use in fouled systems. The results of the biocide challenge (Tables 2 and 3) show that both biocides were quite effective at sharply reducing, and possibly eliminating, viable L. pneumophila cells from the bulk fluid (planktonic population). Bronopol appeared to be more effective than Kathon, at the concentrations recommended by the manufacturers. Since the effective concentration of biocide within a cooling tower falls off rapidly as a result of loss through drift and dilution through the addition of make-up water, the more rapidly a biocide works the more effective it is likely to be. Both biocides were also effective in reducing the number of viable, adherent bacteria. While the results indicated that the biocides were capable of reducing the numbers of sessile bacteria by 8&99% (based on viable counts) within a 9-24 h period, this still left a relatively large population of viable bacteria on the surfaces. T h e presence of this population would serve as an excellent reservoir of bacteria to recontaminate the system after cessation of the biocide dosing. Additionally, the biocide challenges were seen to have little effect on removing the dead bacterial cells from the surfaces as assessed by electron microscopy (micrographs not shown). The presence of bacterial cells or cell products has been shown to promote the more rapid and tenacious recolonization of surfaces in other systems (Nickels et al. 1981). In field studies where the efficacy of various biocides has been tested for their ability to deplete and control L. pneumophila populations, the surfaceassociated mode of growth of L. pneumophila may be one of the factors involved in the poor performance of the biocide as compared with the results expected from laboratory tests where the adherent population of bacteria was not examined. At least two major hypotheses have been published which help explain the generally noted differences in the antimicrobial susceptibilities between adherent and planktonic bacteria. Costerton (1984) proposed that the resistance to antimicrobial killing of sessile bacteria (vs their planktonic counterparts) arises by virtue of their ability to grow adherent to surfaces in glycocalyx-enclosed microcolonies. Brown et al. (1988) also proposed that the increased resistance of sessile bacteria to antibiotics is a result of their slower growth rate. In this particular case, neither of these hypotheses appear able to explain the results on their own. The situation is probably more complex as a result of the bacteria growing in glycocalyxenclosed microcolonies as well as both populations of bacteria growing at a very slow rate as shown by the need for replenishing the system to promote colonization and growth. In a practical sense, the results indicate that the sessile population must be taken into account when evaluating the
eficacy of biocides against cooling tower Legionella spp. At the same time particular attention must be paid to the test conditions so that the experimental system reflects the conditions under which the biocide will be functioning. More experimental work will be required to elucidate the mechanism by which sessile bacteria are afforded greater resistance to the decontamination procedures which are effective against the planktonic bacteria. These types of studies will be best addressed utilizing the carefully controlled environments described by Keevil (1989) for the growth of biofilms. ACKNOWLEDGEMENTS T h e authors wish to thank Glynnis Broadbent and Urvashi Sabharwal for their expert assistance with photography and electron microscopy, and to Bird Archer Co. and Rohm and Haas Co. for the supply of biocides and technical information. This work was supported by grants from the Alberta Heritage Foundation for Medical Research. REFERENCES A L B R E C H T S E N , O., BARTLETT, C.L.R., B E T T E L H E I M , K.A., B O R N S T E I N., N , C O L B O U R NJE. S, . , F L E U R E T T E ,
J . , F R A S E R ,D.W., J O L Y , J.R., K A L L I N G S I, . , M A R T I N - B O U R G NC., O N ,PROSOWSKY, S.B., R O W B O THAM, T.J., STOLWIJK J.A.S., , TARTAKOVSK I .Y S ., , & Y A E U U C H I , E. (1990) Epidemiology, prevention and control of legionellosis : Memorandum from a WHO meeting. Bulletin of the World Health Organisation 68, 155-164.
B R O W NM.R.W., , ALLISON,D.A. & G I L B E R TP. , (1988) Resistance of bacterial biofilms to antibiotics: A growth-rate related effect ? Journal of Antimicrobial Chemotherapy 22, 777778.
C O L B O U R NJE. S, . & D E N N I S ,P.J. (1988) Legionella: A biofilm organism in engineered water systems? In: Biodeterioration 7: Seventh International Biodeterioration Symposium. ed. Houghton, D.R. et al. pp. 3 H 2 . Barking: Elsevier Applied Science. CORDES,L.G., F R A S E RD.W., , S K A L I YP., , PERLINO, C.A., ELSEA,W.R., M A L L I S O NG,. F . & H A Y E SP.S. , (1980) Legionnaires’ disease outbreak at an Atlanta, Georgia country club: Evidence for spread from an evaporative condenser. American Journal of Epidemiology 111,425-53 1. COSTERTON, J.W., I R V I NR , . T . & C H E N G K.-J. , (1981) The bacterial glycocalyx in nature and disease. Annual Reviews of Microbiology 35, 299-324.
COSTERTON, J.W. (1984) The formation of biocide-resistant biofilms in industrial, natural and medical systems. Developments in Industrial Microbiology 25, 363-372.
D O N D E R OT, .J . , R E N D T R O FR.C., F, M A L L I S O NG, . F . , WEEKS,R . M . , L E V Y J, . S . , WONG,E.W. & S C H A F F N E R , W. (1980) An outbreak of Legionnaires’ disease associated with a contaminated air-conditioning cooling tower. New England Journal of Medicine 302, 365-370.
BlOClDE SUSCEPTIBILITY OF L. PNEUMOPHILA 537
E N G L A N 111, D A.C., F R A S E R D ,. W . , M A L L I S O NG, . F . , M A C K E LD.C., , S K A L I YP. , & G O R M A NG, . W . (1983) Failure of Legionella pneumophila sensitivities to predict culture results from disinfectant-treated air-conditioning cooling towers. Applied and Environmental Microbiology 43,240-244. F L I E R M A N C.B. S, & HARVEY R,. S . (1984) Effectiveness of I-bromo-3-chlor0-5,Sdimethylhydantoin against Legionella pneumophila in a cooling tower. Applied and Environmental Microbiology 47, 1307-1310. G A R B E ,P.L., D A V I S ,B.J., WEISFELD,J . S . , M A R K O W I T Z , L., M I N E R ,P . , G A R R I T YF, . , B A R B A R E JE.,M . & REINGOLD, A.L. (1985) Nosocomial Legionnaires’ disease : Epidemiologic demonstration of cooling towers as a source. Journal of the American Medical Association 254, 521-
524. G L I C K ,T . H . , G R E G G ,M.B., B E R M A NB., , MALLISON, G., RHODES,W . W . & KASSANOFF, L . (1978) Pontiac fever. An epidemic of unknown etiology in a health department. I. Clinical and epidemiological aspects. American Journal of Epidemiology 107, 14%160. G R A C ER.D., , D E W A RN.E., , B A R N E SW , . G . & HODGES, G. R. (1981) Susceptibility of Legionella pneumophila to three cooling tower microbiocides. Applied and Environmental Microbiology 41, 233-236. HELMS, C . M . , M A S S A N A R IR, . M . , W E N Z E L , R.P., PFELLER,M.A., M O Y E R , N . P . , H U L L , N . & THE L E GI o N E L L A M O NI T O R I N G C o M M I T T E E ( 1988) Legionnaires’ disease associated with a hospital water system: A fiveyear progress report on continuous hyperchlorination. Journal of the American Medical Association 259, 2423-2427. HORWITZ, W. ( e d . ) (1980) Oficial Methods of Analysis of the Association of Oflcial Analytical Chemists. 4.027-4.028. Washington, DC : Association of Official Analytical Chemists. J O L Y , J . R . & W I N N ,W . C . J R (1984)Correlation of subtypes of Legionella pneumophila defined by monoclonal antibodies with epidemiological classification of cases and environmental sources. Journal of Infectious Diseases 150,667-67 1. K A U F M A N NA , . F . , M C D A D E , J . E . , P A T T O N , C.M., B E N N E T TJ.V., , S K A L I Y P, . , F E E L E Y J, . C . , A N D E R SON,D . C . , P O T T E RM , . E . , N E W H O U S EV, . F . , G R E G G , M.B. & B R A C H M A PN.,S . (1981)Pontiac fever: Isolation of the etiologic agent (Legionella pneumophila) and demonstration of its mode of transmission. American Journa! of Epidemiomology 114, 337-347. K E E V I L ,C . W. (1989) Continuous culture studies of Legionella pneumophila. Journal of Applied Bacteriology 67, x. K E E V I LC.W., , M A C K E R N E SC.W. S, & C O L B O U R N JE.,S . (1990) Biocide treatment of biofilms. International Biodeterioration 26, 169-179. L U F T , J . H . (1971) Ruthenium red and ruthenium violet. I. Chemistry, purification, methods of use for electron microscopy and mechanism of action. Anatomical Record 171, 347-368. M A N G I O N EE.J. , & BROOME,C . V . (1985) Legionellosis: Epidemiology. In Legionellosis ed. Katz, S.M. pp. 6tL73. Boca Raton: Chemical Rubber Company Press. M c C o y , W . F . , W I R E M A NJ ., W . & L A S H E NE, . S . (1986) Efficacy of methychloro/methylisothiazolone biocide against
Legionella pneumophila in cooling tower water. Journal of Industrial Microbiology 1, 4%56. M C D A D E J, . E . , S H E P A R DC, . C . , F R A S E R D,. W . , TSAI, T . R . , REDUS,M . A . , DOWDLE, W.R. & T H EL A B O R ATORY INVESTIGATIO TNE A M (1977)Legionnaires’ disease: Isolation of a bacterium and demonstration of its role in other respiratory disease. New England Journal of Medicine 297,
1 197- 1203. M U R A C AP., , S T O U T J.W. , & Y u , V . L . (1987)Comparative assessment of chlorine, heat, ozone, and UV light for killing Legionella pneumophila within a model plumbing system. Applied and Environmental Microbiology 53,447453. N I C K E L S ,J . S . , BOBBIE, R . J . , L O T T , D.F., M A R T Z , R . F . , BENSON,P.H. & W H I T E , D . C . (1981) Effect of manual brush cleaning on biomass and community structure of microfouling film formed on aluminum and titanium surfaces exposed to rapidly flowing seawater. Applied and Environmental Microbiology 41, 1442-1453. P A G E ,S . & G A Y L A R D EC., (1990) Biocide activity against Legionella and Pseudomonas. International Biodeterioration 26,
139-148. POPE, D . H . , S O R A C C OR.J., , GILL, H.K. & FLIERM A N S , C.B. (1982)Growth of Legionella pneumophila in twomembered cultures with green algae and cyanobacteria. Current Microbiology 7 , 319-322. RUSESKA, I., R O B B I N SJ,. , L A S H E NE ,. S . & COSTERTON, J .W . (1982) Biocide testing against corrosion-causing oilfield bacteria helps control plugging. Oil and Gas Journal 8, 253-
264. S C H O F I E LGD. M , . & L O C C I ,R. (1985)Colonization of components of a model hot water system by Legionella pneumophila. Journal of Applied Bacteriology 58, 151-162. S T A T E S ,S.J., C O N L E Y ,L . F . , K U C H T A J, . M . , O L E C K , B.M., L I P O V I C HM.J., , WOLFORD,R.S., W A D O W S K Y , R.M., MCNAMARA A.M., , S Y K O RJA. L, . , K E L E T I ,G . & Y E E , R. B . (1987a)Survival and multiplication of Legionella pneumophila in municipal drinking water systems. Applied and Environmental Microbiology 53, 979-986. STATES, S . J . , C O N L E Y , L . F . , T O W N E R , S . G . , WOLFORD, S . , STEPHENSON, T.E., MCNAMARA, A.M., W A D O W S K Y R,. M . & YEE, R.B. (1987b) An alkaline approach to treating cooling towers for control of Legionella pneumophila. Applied and Environmental Microbiology 53, 1775-
1779. S T O U TJ, . E . , BEST, M . G . & Yu, V . L . (1986) Susceptibility of members of the family Legionellaceae to thermal stress : Implications for heat eradication methods in water distribution systems. Applied and Environmental Microbiology 52, 396-399. S W A N G OL.J., , W I L T , G . R . , K I L L E NA , .D., WILLIAMS, D . E . & W O R L E Y ,S . D . (1987) Inactivation of Legionella pneumophila by hypochlorite and an organic chloramine. Applied and Environmental Microbiology 53, 2983-2986. S Y K E S ,J . M . & B R A Z I E R A , . M . (1988) Assessment and control of risks from Legionnaires’ disease. Annals of Occupational Hygiene 32, 63-67. T H A C K ESR . B, . , B E N N E T TJ, . V . , T S A I , T . F . , F R A S E R , D . W . , M C D A D E ,J . E . , S H E P A R DC , .C., WILLIAMS,
538 J . B . W R I G H T E r A L .
K . H . J R , S T U A R TW , .H., DULL, H.B. & EICKHOFF, T. C . (1978) An outbreak in 1965 of severe respiratory illness caused by the Legionnaires’ disease bacterium. Journal of Infectious Diseases 138, 512-519. TIMBURY M,. C . , D O N A L D S O NJ, . R . , M C C A R T N E Y , A.C., F A L L O NR.J., , S L E I G H , J . D . , L Y O N , D., O R A N G EG.V., , B A I R D D.R., , W I N T E R J, . & W I L S O N , T . S . (1986) Oubreak of Legionnaires’ disease in Glasgow Royal Infirmary : Microbiological aspects. Journal of Hygiene 97, 393403. V I C K E R SR.M., , Yu, V.L., H A N N A ,S., M U R A C AP, . , D I V E N W., , C A R M E NN, . & T A Y L OFR. B , . (1987) Determinants of Legionella pneumophila contamination of water distribution systems : 15-hospital prospective study. Infection Control, 8, 357-363. W R I G H TJ.B. , & COSTERTON J .,W . (1986) Visualization of the glycocalyx of Legionella pneumophila. Abstracts of the Annual Meeting of the Alberta Heritage Foundation f o r Medical Research p. 26. Edmonton: Alberta Heritage Foundation for Medical Research.
W R I G H T ,J . B . , A T H A R , M . A . , V A N O L M ,T . M . , WOOTL I F F , J . S . 8~ C O S T E R T O N J ., W . (1988) Atypical legionellosis: Isolation of Legionella pneumophila serogroup 1 from a patient with aspiration pneumonia. Journal of Hospital Infection 13, 187-190. W R I G H T A.E., , B A R T L E T TC, . L . R . , B R U N D R E TG T ,. W . , H I L L , E.C., H U G O , W.B., H U T C H I S O N J, . G . P . , J O N E S ,W . P . , L E E , J., P I K E , E . B . , S P O O N E RD, . F . , TILLETT H ,. E . & P E N N ,R . G . (1989a) Report ofthe Expert Advisory Committee on Biocides. Department of Health, London : HMSO. W R I G H T ,J . B . , RUSESKA,I . , A T H A R , M.A., C O R B E T T , S. & COSTERTON,J.W. (1989b) Legionella pneumophila grows adherent to surfaces in vitro and in situ. Infection Control and Hospital Epidemiology 10, 4 0 8 4 1 5 . YAMAMOTO H,. , U R A K A M I ., , N A K O N OK , . , I K E D O ,M . & Y A B U U C H IE, . (1987) Effect of FLONLIZER ultraviolet sterilizer on Legionella species inhabiting cooling tower water. Microbiology and Immunology 31,745-752.
