Effect of Growth in Biofilms on Chlorine Susceptibility of Mycobacterium avium and Mycobacterium intracellulare

Abstract

Mycobacterium avium and Mycobacterium intracellulare were grown in suspension and in biofilms, and their susceptibilities to chlorine were measured. M. avium and M. intracellulare readily adhered within 2 h, and numbers increased 10-fold in 30 days at room temperature in biofilms on both polystyrene flasks and glass beads. The chlorine resistance of M. avium and M. intracellulare cells grown and exposed to chlorine in biofilms was significantly higher than that of cells grown in suspension. Survival curves showed no evidence of a resistant, persisting population after 6 h of exposure to 1 μg chlorine/ml. The chlorine susceptibility of cells grown in biofilms and exposed in suspension (cells detached from bead surfaces) was also significantly higher than that of cells grown and exposed in suspension (planktonic cells), although it was lower than that of cells grown and exposed in biofilms. The higher resistance of the detached biofilm-grown cells was reversed upon their growth in suspension. There was a strong correlation between the chlorine susceptibility of cells of both M. avium and M. intracellulare and cell surface hydrophobicity measured by contact angle for both biofilm- and suspension-grown cells.

Mycobacterium avium and Mycobacterium intracellulare are opportunistic environmental pathogens (17, 52). One of the sources of M. avium for human infection is drinking water (50). M. avium and M. intracellulare can be recovered from water (6, 11, 15, 16, 21, 50) and biofilm (19, 46) samples from drinking water distribution systems. M. intracellulare numbers in biofilms averaged 600 CFU per cm² (19), suggesting that this mycobacterium may prefer the biofilm habitat. M. avium has also been reported as colonizing and growing in point-of-use water filters containing silver as a disinfecting agent (37). Other mycobacteria, including Mycobacterium kansasii (38), Mycobacterium flavescens (39), Mycobacterium chelonae (40), Mycobacterium fortuitum (23), and Mycobacterium phlei (1), have also been shown to form biofilms.

The high cell surface hydrophobicity of M. avium and M. intracellulare is likely to contribute to the formation of biofilms (9). In fact, the adhesive properties (e.g., hydrophobicity) of mycobacteria, corynebacteria, and nocardiae are related to the chain length of their mycolic acids (3). M. avium and M. intracellulare are concentrated at air-water interfaces (51) and adhere to particulate matter in soil (5) and water (19). Strains of those species are also relatively resistant to heavy metals (20). Heavy-metal resistance (i.e., Ag) of mycobacteria undoubtedly contributed to the colonization of point-of-use water filters by M. avium (37). M. avium and M. intracellulare are also highly resistant to chlorine and other disinfectants used in water treatment (6, 18, 35, 45). Chlorine exposure of microbial biofilms on copper pipes in a model distribution system led to loss of almost all microorganisms except mycobacteria (34). The combination of high hydrophobicity and resistance to chlorine and heavy metals suggests that M. avium and M. intracellulare are biofilm pioneers in drinking water distribution systems, especially on copper and galvanized pipes.

Cells of M. avium and M. intracellulare in biofilms may also be more resistant to antimicrobial agents. A number of mechanisms for enhanced resistance to antimicrobials of biofilm-grown cells have been proposed (29). First, layers of cells and extracellular material comprising the biofilm (25, 33) could provide protection because penetration of chlorine is reduced in biofilms (8, 12). Second, physiological changes in cells could result from growth in the biofilm habitat (14, 24, 27, 28, 29). Cells of M. avium grown at low growth rates, through changes in nutrient composition, exhibited increased chlorine resistance (18). Cells in biofilms may exhibit changes, such as reduced growth rates (14), as a result of reduced oxygen (13, 47, 53) and nutrient penetration in biofilms (42). Third, the frequency of physiologically resistant cells (“persisters”) could be higher in biofilm populations (26, 29). Fourth, high concentrations of organic matter in biofilms would increase chlorine demand, thereby making cells appear chlorine resistant.

Because of the potential significance of biofilm formation to persistence in drinking water distribution systems in the presence of residual disinfectant (e.g., chlorine), we initiated a study of the characteristics of M. avium and M. intracellulare in biofilms. Here we report the chlorine susceptibility of cells in suspension, of cells in biofilms, and of cells detached from biofilms. In addition, we report the relationship between the chlorine susceptibility and hydrophobicity of those different cell types.

MATERIALS AND METHODS

Mycobacterial strains.

M. avium strain A5 (2) and M. intracellulare strain TMC 1406^T (31) were used.

Growth of strains.

Single isolated colonies on Middlebrook 7H10 agar medium (BBL Microbiology Systems, Cockeysville, MD) containing 0.5% (vol/vol) glycerol and 10% (vol/vol) oleic acid-albumin (36) were used to inoculate 2 ml of Middlebrook 7H9 broth medium (BBL Microbiology Systems, Cockeysville, MD) containing 0.5% (vol/vol) glycerol and 10% (vol/vol) oleic acid-albumin (36). After 7 days of incubation at 30°C, 1 ml was used to inoculate 9 ml of M7H9 or sterilized potable water obtained from the New River Valley Municipal Water System (assimilable organic carbon range, 500 to 750 μg/liter). Those stock cultures were incubated for 7 days at 30°C and used to inoculate 40 ml M7H9 in polystyrene flasks (Falcon, 50 ml, canted neck, 0.2-μm vented plug seal cap; Becton Dickinson, Franklin Lakes, NJ). To grow mycobacterial strains in biofilms on the surfaces of polystyrene flasks or on 4.0-mm-diameter glass beads (Fisher Scientific, Pittsburgh, PA), cells from stock cultures were inoculated into 40 ml M7H9 in polystyrene flasks containing glass beads. Cultures were incubated at room temperature and slowly rotated at 100 rpm. Unless otherwise noted, cells in biofilms were tested after 30 days of growth. For some experiments, the potable water was not sterilized and was used as the growth medium (with 10% M7H9) to measure the effect of the presence of the water's normal microbial flora on mycobacterial biofilm formation.

Measurement of biofilm accumulation.

To measure the number of CFU from glass bead biofilms (unexposed or chlorine exposed), a specified number of beads was transferred from a flask using a sterile loop to a petri dish containing 25 ml chlorine demand-free water (CDFW) (10). The beads were washed by swirling at 30 rpm for 30 s. The beads were transferred to a tube containing 0.8 ml CDFW, 0.1 ml 0.1% (wt/vol) sodium thiosulfate, and 0.1 ml of 0.05% (vol/vol) Tween 80. The cells were separated from the beads by vortexing for 1 min at the highest speed. The turbidity of the suspension was measured and 0.1 ml of the suspension and dilutions in CDFW broth spread on M7H10. Colonies were counted after 7 days at 30°C.

Measurement of chlorine susceptibility.

To measure the chlorine susceptibility of biofilm-grown cells that were detached from biofilms on bead surfaces, beads were transferred to 25 ml CDFW in a petri dish and gently washed as described above. Five washed beads were transferred into 1 ml of CDFW, and cells were detached by vortexing at the highest speed for 1 min. The beads were allowed to settle, and the supernatant suspension was used (biofilm-grown, liberated cells). Suspensions of mycobacterial cells freed of aggregates of cells (i.e., reduced aggregate fractions) (45) were prepared from cultures grown in suspension and from suspensions of cells liberated from biofilms before chlorine exposure to reduce the possible protective effect of cells in aggregates. Suspensions were diluted to 10⁴ CFU/ml in 25 ml CDFW in a foil-covered flask, and the colony count was measured by spreading 0.1 ml of the suspension and a 10-fold dilution on M7H10 agar. Chlorine was then added to a final concentration of 1 μg/ml (1 ppm), and 1.8-ml samples were removed immediately and at 1, 2, and 3 h of incubation at room temperature and then mixed with 0.2 ml 0.1% (wt/vol) sodium thiosulfate to inactivate the chlorine. Undiluted and 10-fold-diluted samples in CDFW (0.1 ml) were spread on M7H10 agar, and surviving colonies were counted after 7 to 10 days of incubation at 37°C.

The chlorine susceptibility of cells in intact biofilms was measured on cells harvested from glass beads at different times following addition of chlorine. Beads were transferred from polystyrene flasks with a sterile loop and washed in 25 ml of CDFW in a sterile petri dish by gentle rotation as described above. The washed beads were then transferred to 25 ml CDFW in a foil-covered flask. Immediately five beads were removed and placed in CDFW containing sodium thiosulfate and Tween 80, and the CFU was measured as described above. Chlorine was added to a final concentration of 1 μg/ml, and at 10, 20, 40, and 60 min, five beads were removed and the surviving CFU of cells on the beads was measured as described above for determination of growth rate. The presence of either sodium thiosulfate or Tween 80 at the concentrations used did not influence survival.

To follow the disappearance of the adaptive resistance of biofilm-grown cells of M. avium strain A5, biofilm cultures grown 30 days at room temperature on glass beads were washed in 25 ml of CDFW in a petri dish. Five beads were transferred to 1 ml of CDFW, and cells were liberated by vortexing at the highest setting for 1 min. The beads were allowed to settle, and 0.9 ml of liberated cells was added to 21.6 ml of CDFW and 2.5 ml of M7H9 broth in a 100-ml flask and incubated at room temperature. Immediately, and at daily intervals up to 1 week, reduced aggregate fractions of the cells were prepared and chlorine susceptibility was measured.

The data represent the average of a minimum of three replicates. CFU counts were used to calculate the logarithm of the percent survival with time (in minutes) for each strain and condition of growth. CT_99.9 values were calculated by multiplying the concentration of chlorine (in parts per million) by the time (duration of exposure, in minutes) for a 99.9% reduction in the CFU count.

Measurement of chlorine concentration.

Chlorine concentration was measured before and during exposure using the N,N-diethyl-p-phenylenediamine Free Chlorine reagent (catalog no. 21055-28; Hach Chemical Co., Loveland, CO). The presence of 10% M7H9 did not alter the standard curve of chlorine.

Contact angle measurement.

To measure the number of CFU and harvest cells from biofilms on polystyrene, cells were scraped from the entire bottom surface using a Teflon scraper and suspended in 10 ml CDFW. The suspension was vortexed at maximum speed for 60 s to prepare a uniform suspension for measurement of turbidity and CFU per milliliter as described above. Cells grown in suspension or detached from biofilms were collected on 0.45-μm-pore-size, 25-mm-diameter cellulose acetate filters (Millipore, Bedford, MA) that had been soaked in a glycerol-water (50:50) solution. The four quadrants of the filters were glued to the surface of a glass slide, and the contact angles of 20-μl droplets of water, n,n-dimethylformamide, and hexadecane were measured with a goniometer (49). The contact angles of cells in intact biofilms were measured on polystyrene following incubation of cells. The culture was removed from the flasks, the flask contents were dried, and the wall of the flask with the biofilm was cut away. Droplets were placed on the biofilm, and the contact angles were measured with a goniometer.

Statistical analysis.

Statistical analysis (e.g., t test and correlation) was performed using InStat, version 3.0 (GraphPad Software, Inc., San Diego, CA).

RESULTS

Rates of adherence and growth in biofilms.

Cells of M. avium strain A5 and M. intracellulare strain TMC 1406^T readily adhered to both polystyrene and glass surfaces whether grown in M7H9 medium or water. Although the inoculated water-grown cultures contained only 5 × 10⁶ CFU/ml, glass beads in water collected 2 h after inoculation at room temperature had approximately 2,000 CFU/cm² of either M. avium strain A5 or M. intracellulare strain TMC 1406^T (Table 1). Those values provide a reliable estimate of the extent of adherence, because growth of the slowly growing mycobacteria should not have affected the results.

TABLE 1.

Adherence and accumulation of cells of M. avium strain A5 and M. intracellulare TMC 1406^T to glass beads in water^a

Incubation periodb	Mean CFU/cm2 of bead (±SD)c
	M. avium A5	M. intracellulare TMC 1406T
2 h (adherence)	2.0 × 103 (±0.31)	1.8 × 103 (±0.12)
1 wk	1.6 × 103 (±0.10)	1.6 × 103 (±0.50)
2 wk	3.9 × 103 (±0.20)	1.0 × 103 (±0.32)
3 wk	5.2 × 103 (±0.09)	1.9 × 103 (±0.35)
4 wk	2.9 × 104 (±0.68)	2.5 × 104 (±0.34)

^aWater cultures contained 10% M7H9 broth.

^bTime after inoculation of glass bead-containing water medium incubated at room temperature.

^cn = 3.

Numbers of both M. avium strain A5 and M. intracellulare strain TMC 1406^T on the surfaces of the glass beads increased approximately 10-fold by 4 weeks of incubation at room temperature in water (Table 1). Microscopic examination revealed that the extent of the surface covered by microbial cells had increased by 4 weeks of incubation to the point where the entire glass or polystyrene surface appeared to be covered with mycobacterial cells (41). There was a considerable lag before colony numbers increased in the glass bead-water cultures. Using the data from 3 to 4 weeks, the doubling time for M. avium strain A5 on the glass beads was 2.8 days and that for M. intracellulare strain TMC 1406^T was 2.1 days. It should be pointed out that the data in Table 1 for CFU per square centimeter and the calculated rates of growth shown above represent the rates of accumulation of cells on the surfaces of glass beads, not true growth rates. Because the cells in the biofilms were enumerated from beads in flasks containing cells growing in suspension (planktonic) as well as in the biofilms, the term “accumulation” is a more accurate representation of the measurement. The increase in the number of cells in the biofilms was likely due to growth of adherent cells plus the binding of new cells to the biofilm minus the cells spontaneously detached from the biofilms. The increases in numbers of mycobacteria on beads in the water cultures were expected because the water had an assimilable organic carbon level of approximately 500 μg/liter and contained 1/10-strength M7H9 medium from the inoculum. Further, both M. avium and M. intracellulare can grow in natural and drinking waters (19, 22) and above an assimilable organic carbon level of ≥50 μg/liter (34). The numbers of CFU per square centimeter in the experiments here are comparable to the numbers in biofilms in drinking water systems (19).

Separate experiments were performed with biofilms produced after inoculation of M. avium strain A5 into nonsterile water containing glass beads and incubation for 30 days at room temperature. The presence of the water's normal microbial flora in biofilms reduced the number of M. avium bacteria by a factor of 3, from 3.0 × 10⁴/cm² (M. avium strain A5 alone) to 1.0 × 10⁴/cm² (M. avium strain A5 plus normal flora).

Vortexing the glass beads removed almost all the cells from the beads as judged by acid-fast microscopy of beads stained after vortexing. There was little bead-to-bead variation in the number of cells recovered from beads. In two separate experiments, glass bead biofilms contained 384 ± 96 and 342 ± 111 M. avium strain A5 CFU/cm², and the differences were not significant. Exposure of cells to 0.005% Tween 80 or sodium thiosulfate and vortexing, as was performed for recovery of cells from glass bead biofilms, did not change the number of CFU compared to vortexing in water alone (data not shown), in agreement with the findings of Naik et al. (32).

Increased chlorine resistance of cells in biofilms and cells harvested from biofilms.

The chlorine susceptibilities of the three cell populations (sessile, detached, and planktonic) of M. avium strain A5 and M. intracellulare strain TMC 1406^T are displayed in Table 2. Chlorine susceptibility is expressed as CT_99.9, the product of chlorine concentration in micrograms per milliliter (parts per million) and duration of chlorine exposure (in minutes) required to kill 99.9% of cells (as CFU). Cells were grown on glass beads in water plus 10% M7H9 for the experiments reported, but similar results were obtained for cells grown on polystyrene (41). Cells of M. avium strain A5 and M. intracellulare strain TMC 1406^T grown in suspension (planktonic) and exposed in suspension were the most susceptible. Cells grown in biofilms (sessile) and exposed in biofilms were the most resistant. Cells of either strain grown in biofilms but detached from beads and exposed to chlorine in suspension displayed intermediate susceptibility (Table 2). Chlorine concentrations remained within the range of 1 ± 0.1 ppm throughout the duration of the experiments.

TABLE 2.

Chlorine susceptibility of cells of M. avium and M. intracellulare grown in M7H9 medium or water in suspension or in biofilms

Species and growth medium	CT99.9 (mean ± SD)a for the following cell populationb:
	Planktonic	Detached	Sessile
M. avium strain A5
M7H9	150 ± 75c	358 ± 173c	417 ± 249c
Water	209 ± 71	356 ± 162c	835 ± 236c
M. intracellulare strain     TMC 1406T
M7H9	380 ± 183	458 ± 248	607 ± 322c
Water	330 ± 166	435 ± 228	593 ± 477c

^aFor four independent triplicate counts. CT_99.9 is calculated as the product of the chlorine concentration (in parts per million) and the duration of exposure (in minutes) for a 99.9% reduction in the colony count.

^bPlanktonic, cells grown in suspension and exposed in suspension; detached, cells grown in biofilms and exposed in suspension; sessile, cells grown in biofilms and exposed in biofilms.

^cSignificantly different (P <0.05 by Student's t test) from other values in the same row.

Separate experiments employed the water without sterilization to permit the formation of a mixed biofilm (i.e., M. avium and the normal flora). Following establishment of the biofilm (30 days at room temperature), the biofilms were exposed to chlorine, and survival was measured as described above. The presence of the water's normal microbial biofilm increased the susceptibility of biofilm-grown cells of M. avium in the biofilm. The survival (expressed as a percentage of initial CFU) of M. avium cells grown alone in biofilms was 1.6 times higher than that of M. avium cells grown in the presence of the normal biofilm flora.

Because the chlorine resistance of all three populations of cells was so high (at least 300-fold higher than that of Escherichia coli [45]), we sought to determine whether there existed a resistant, persisting population (26, 29). Survival measurements were carried out with higher numbers of beads and up to 6 h exposure to 1 μg chlorine/ml. Chlorine concentrations remained at 1 μg chlorine/ml for the full 6-h exposure. There was no evidence of a resistant or “persister” (26, 29) population. The survival curves for cells of M. avium strain A5 and M. intracellulare strain TMC 1406^T were linear from 0 to 6 h. The extreme chlorine resistance of these mycobacteria is illustrated by the fact that approximately 5% of cells survived after a 6-h exposure to 1 μg chlorine/ml.

Loss of chlorine resistance by detached, biofilm-grown cells.

The increased chlorine resistance of cells of M. avium strain A5 and M. intracellulare TMC 1406^T grown in biofilms in water, detached from bead surfaces, and exposed to chlorine in suspension was reversed by growth of cells in suspension. The increased chlorine resistance of the detached cells was lost following a 1-day incubation in suspension for both M. avium strain A5 (CT_99.9, 330 ± 125 to 106 ± 40) and M. intracellulare strain TMC 1406^T (CT_99.9, 440 ± 64 to 122 ± 46).

Relationship between hydrophobicity and antimicrobial susceptibility.

Cell surface hydrophobicity, measured by the contact angle of cells grown on polystyrene (Table 3), was correlated with survival following exposure to chlorine for both M. avium and M. intracellulare whether grown in M7H9 medium or water (Table 4). Cells grown on polystyrene, detached from that surface, and exposed to chlorine in suspension exhibited the same CT_99.9 values as did cells grown on glass beads (41) (data not shown). The correlation coefficients ranged from a low of 0.781 to 0.997, demonstrating that cell surface hydrophobicity was a major determinant of the chlorine susceptibility of both M. avium strain A5 and M. intracellulare strain TMC 1406^T.

TABLE 3.

Hydrophobicity^a of M. avium strain A5 and M. intracellulare strain TMC 1406^T

Growth and medium	Cosine of contact angle (°)
	M. avium A5	M. intracellulare 1406T
Suspension
M7H9	0.927	0.840
Water	0.883	0.826
Detached from biofilms
M7H9	0.883	0.716
Water	0.798	0.588
Intact biofilms
M7H9	0.669	0.569
Water	0.602	0.340

^aExpressed as cosine of contact angle.

TABLE 4.

Correlation between cosine of contact angle and chlorine CT_99.9 values

Growth medium	Correlation coefficienta
	M. avium A5	M. intracellulare 1406T
M7H9	0.781	0.992
Water	0.997	0.987

^aFor correlation between chlorine CT_99.9 and paired cosine of contact angle of water on cells. Three paired values were used for each data point.

DISCUSSION

M. avium and M. intracellulare, like M. kansasii (38), M. flavescens (39), M. fortuitum (23), and M. phlei (1), readily form biofilms. Whether adherence to the glass or polystyrene required the presence of proteins or other compounds was not determined. Biofilm formation of M7H9-grown cells of a variety of M. avium strains, including A5, in water and M7H9 was dependent on Ca²⁺, Mg²⁺, and Zn²⁺ (7). However, adherence independent of growth was not measured (7). It is possible that the hydrophobicity of mycobacterial cells is all that is needed for adherence.

The results demonstrate that the chlorine resistance of M. avium and M. intracellulare cells in biofilms is significantly increased over that of cells grown in suspension (Table 2). Those data agree with the disinfectant resistance of biofilm-grown M. phlei (1). Further, cells detached from glass bead biofilms and exposed to chlorine in suspension were more resistant than cells grown in suspension (Table 2). Although Tween 80 did not reduce the CFU of unexposed cells, it might have increased the chlorine susceptibility of the biofilm-grown cells (18, 32), even considering the short duration of exposure and the fact that chlorine was inactivated by sodium thiosulfate when Tween 80 was added. Thus, the chlorine survival and CT_99.9 values reported here for the biofilm-grown cells exposed to chlorine while in biofilms might be underestimates.

It is likely that growth rate differences influenced the chlorine susceptibility of biofilm-grown cells measured, as they do for cells in suspension (18). Growth, independent of adherence of cells to surfaces or liberation of cells from surfaces, was not measured. The reduced growth rate of cells in biofilms has been cited as a factor contributing to antibiotic resistance (14). Slowly growing cells of M. avium and M. intracellulare are more resistant to chlorine than rapidly growing cells (18).

Based on these results, it appears that two factors influence the chlorine susceptibility of M. avium and M. intracellulare cells in biofilms. One factor is likely the limited diffusion and penetration of chlorine (8, 12) resulting from the presence of layers of cells and extracellular material (25, 33). That is illustrated by the fact that the highest CT_99.9 values were measured in cells grown and exposed to chlorine in biofilms. Unfortunately, there has been no description of the composition of mycobacterial biofilms to date. A second factor is necessary to explain the adaptive, increased resistance of biofilm-grown cells exposed to chlorine in suspension. For those liberated cells, there was no extracellular matrix protecting the cells, because of the washing steps involved in their preparation and dilution before exposure to chlorine. Possibly, physiological changes in cells as a consequence of the biofilm habitat altered their susceptibility to chlorine.

Because the chlorine susceptibilities of all three types of cells were measured here, it was possible to identify the increased resistance of cells grown in biofilms but liberated from them. Although Schulze-Röbbecke and Fischeder (39) measured the chlorine susceptibilities of cells of M. kansasii and M. flavescens grown and harvested from biofilms (i.e., liberated), they compared the susceptibilities only to those of cells grown and exposed in biofilms. The chlorine susceptibility of suspension-grown cells was not reported (39). Thus, the data did not permit identification of the increased and adaptive chlorine resistance of biofilm-grown cells. Physiological changes of other microorganisms as a consequence of growth in biofilms have been postulated to result in increased antimicrobial resistance of biofilm-grown cells (24, 27, 28, 29). Although oxygen availability is reduced for cells in biofilms (13, 47, 53), it is not likely a factor here, because growth of M. avium cells in suspension at 6% oxygen resulted in increased chlorine susceptibility (18).

One physiological characteristic influencing the chlorine susceptibility of M. avium and M. intracellulare was cell surface hydrophobicity. It has been well established that mycobacteria are among the most hydrophobic of microorganisms (49), and hydrophobicity has been shown to promote surface adherence (48). Chlorine susceptibility, expressed as CT_99.9 values, correlated with hydrophobicity, as reflected by the cosine of contact angle (Tables 3 and 4). It is possible that M. avium and M. intracellulare change the cell surface lipid and fatty acid composition in response to whether cells are in suspension or on surfaces. Mycobacteria have been shown to alter fatty acid composition in response to temperature differences (43, 44). Because the mycolic acid structure is a determinant of membrane fluidity in mycobacteria (4, 30), alterations in mycolic acids would be expected to change membrane permeability to chlorine.

The discovery that M. avium and M. intracellulare cells liberated from biofilms are transiently more resistant to chlorine has important ramifications for water treatment. First, the CT_99.9 values of suspension-grown and -exposed M. avium and M. intracellulare are significantly lower than those of cells in biofilms or released from biofilms. Thus, such CT_99.9 values are misleading. Not only are cells in biofilms more chlorine resistant, but cells freed from biofilms can survive and adhere to other pipe locations to initiate biofilm formation downstream.