Laboratory Scale Model System

This study was conducted using a model cooling tower system (100 l) under constant hydraulic conditions and temperature (37°C). A water heater (AT-100, 12 V, 100 W, Atman, Germany) and a recirculation pump (40 l min⁻¹, 550 W, Pedrollo SpA, Italy) were installed to model system to provide the desired control temperature and facilitate evaporation inside the bulk water. Public drinking water was used to replace water lost by evaporation and partial drainage after neutralising the residual free chlorine (0.2–0.4 mg l⁻¹) with sodium thiosulfate (STS). Throughout the experiment, no chemicals were injected into the model system to prevent their potential adverse impacts on biofilm formation and growth. The pH value, total dissolved solids (TDS), conductivity and dissolved oxygen (DO) content of the water samples were analysed monthly.

For modelling legionellae-contaminated microbial flora in a cooling tower, L. pneumophila ATCC 33152 standard strain (final concentration of 1.2 × 10⁵ colony forming units (CFU) ml⁻¹) and the water of an industrial cooling tower (1:5 ratio, actual cooling system water:potable water) was inoculated to the model system.

Test Slides

Before being inserted into the model system, PVC, stainless steel (AISI 316) and glass slides (25 × 60 × 0.8 mm) were cleaned using 1% Triton and then degreased in acetone for 3 min using an ultrasonic cleaner (220 V, 60 Hz, 600 W, Alex Machine, Turkey). After rinsing with sterile deionised water, the slides were dried at 50°C and sterilised using ultraviolet (UV) light for 12 h (9). Sterile slides were inserted into the channels of the biofilm holder with a 2 cm spacing. Biofilms were allowed to grow for six months on slides in the aqueous phase of the system.

Biocide Preparation

Chloramine T trihydrate was dissolved in sterile deionised water at a final concentration of 500 mg l⁻¹. Free and total chlorine concentrations were quantified using the N,N-diethyl-p-phenylenediamine (DPD) method (Model CN-66, Hach^®, USA). According to the manufacturer’s recommendation, the residual active biocide concentration was calculated by multiplying the combined chlorine concentration (total chlorine and free chlorine) value by 3.97 (21).

Sampling of Water and Biofilms and the Application of Biocide

Biofilm and water samples were collected monthly, the biocide chloramine T trihydrate was applied to each water or biofilm sample for 1 h or 3 h of contact time in a sterile container. After the incubation period, the samples were neutralised with 0.5% STS, and the resulting suspensions were serially diluted in sterilised tap water to 10⁻⁵. The samples were evaluated using a conventional culture method, total or free ATP concentration measurements and total and viable cell counts after fluorescence staining. Each test was repeated three times.

Heterotrophic Bacteria and L. pneumophila Counts

The number of cultivable bacteria in the water or biofilm samples was quantified using the spread plate technique on R2A medium. The colonies on the plates were enumerated after incubating for 7–10 days at 28°C.

The water and biofilm (biocide-exposed and unexposed) samples were also examined for the presence of cultivable L. pneumophila on alpha-ketoglutarate-enriched buffered charcoal yeast extract (BCYE) agar (Oxoid Ltd, UK) by incubating the plates at 37 ± 1°C for up to 10 days according to ISO 11731:2017 standard (22). The results were reported as CFU ml⁻¹ for water samples and CFU cm⁻² for biofilm samples, based on the surface areas of the coupons (30 cm²).

ATP Measurements

Samples were collected using free and total ATP swabs after 1 h or 3 h of contact time. ATP was extracted from the cells with a cationic detergent, after which the nascent ATP was measured after reacting with the luciferin-luciferase system. The swabs were placed into a luminometer (Biotrace^TM Uni-Lite^TM NG, 3M, USA), and the intensity of the light was converted into relative light units (RLU) shortly thereafter (23).

The results were recorded as RLU ml⁻¹ and RLU cm⁻² for the water and biofilm samples, respectively. The microbial ATP concentration was indirectly calculated by determining the difference between total and free ATP levels.

Fluorescence Staining

The number of total and actively respiring bacteria in the samples was assessed by fluorescence staining to examine their responses to the biocide. 5-Cyano-2,3-ditolyltetrazolium chloride (CTC, Sigma-Aldrich, Germany) was utilised as an indicator of ETS-active bacteria (red in colour). The DNA/RNA-specific fluorochrome 4’,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich, Germany) was used as an indicator to obtain the total cell count (blue in colour). The samples were stained with CTC followed by DAPI to enumerate total and respiring cells within the same sample (24). After staining the samples were filtered on black polycarbonate membranes (0.2 μm pore size, MilliporeSigma, Germany) and viewed by epifluorescence microscopy (Eclipse 80i, Nikon Corp, Japan). The results were reported as log cells ml⁻¹ and log cells cm⁻² for the bulk water and biofilm samples, respectively.

Statistical Analysis

The data were analysed using the statistical program SPSS^®, version 22.0 (25). Student’s t-test was performed to compare biocide-treated and untreated samples. The Kruskal-Wallis nonparametric test was used to determine if there were significant differences in the response of the tested biofilms on the different assayed materials under the same conditions. The statistical calculations were based on a 95% confidence level.

Model System

From the first day onward, heterotrophic plate count (HPC), legionellae, pH and TDS values of the water in the model system were higher than those observed in the makeup water, even with a regular blowdown regime (Table I). In this regard, the model cooling tower system used in this study was validated as a potential legionellae-contaminated reference system to investigate its associated physicochemical and microbiological characteristics (Table I).

Heterotrophic Plate Count and Legionella Analysis

The active residual concentration of 500 mg l⁻¹ chloramine T trihydrate in solutions was determined to be 317.6 ± 10. Chloramine T trihydrate was tested as a biocide in this study due to its broad-spectrum efficacy without the risk of inducing bacterial resistance, and it is also noncorrosive to common materials used in industrial applications (21). Chloramine T trihydrate is an organic compound that is produced by chlorinating benzenesulfonamide or para-toluenesulfonamide, dissolving in water and ionising to form the chloramine T ion. The biocidal effect of the chloramine T ion is attributed to its oxidising activity, and it rapidly destroys cell material or disrupts essential cellular processes (26).

A ≥1 log reduction, which is considered a minimum level of performance for biocidal treatment according to ASTM E645-13 guidelines, was observed for planktonic cultivable bacteria for six months (Figure 1(a)) and only for the first two months for the biofilm-associated bacteria (Figures 1(b)–1(d)) (27). However, from the third month onward, the responses of the biofilm samples to the biocide were convoluted due to their cultivation and therefore fluctuated (Figures 1 and 2).

No cultivable planktonic L. pneumophila cells were observed after treatment with the biocide during the six month test period (Figure 1(a)). In contrast, the sessile cultivable heterotrophic bacteria and L. pneumophila responses to the biocide fluctuated monthly. After the first month, the highest log decrease value (>3 log) was determined for biofilms on glass treated with the biocide for 3 h. From the third to the fourth months, with the maturation of biofilms, the sensitivity of the biofilms on all tested materials decreased (≤1 log reduction). In the fifth and sixth months, higher log reduction values were obtained in the sessile HPC counts (>2 log and ≤1 log for the glass or PVC and stainless steel biofilms, respectively) (Figures 1(b)–1(d)). No sessile cultivable L. pneumophila were detected after exposure to the biocide after the third month for all assayed materials.

Variability or fluctuations in the response of both sessile and planktonic cells to biocides can be attributed to biofilm maturation. Biofilm formation and maturation occur in several stages, and due to physiological and metabolic changes that occur at each stage, the antimicrobial susceptibility of the biofilm, and consequently the bulk water flora, can vary (16). The dispersal of cells contributes to biofilm rejuvenation and recalcitrance toward antibiofilm agents (28). After maturation, the sensitivity of the biofilms on all the tested materials decreased (≤1 log reduction). In the fifth and sixth months, greater reductions in the sessile HPC counts were observed again, >2 log and ≤1 log for the glass or PVC and stainless steel biofilms, respectively (Figures 1(b)–1(d)).

ATP Measurements

Because cells injured by sublethal levels of biocides require longer detection times and since biocidal treatment may induce the development of viable but nonculturable (VBNC) cells, the colony counting method results in an underestimation of viable cell numbers. Thus, the number of live or metabolically active bacteria needs to be evaluated by alternative methods, such as ATP measurements or fluorescence staining (10, 29).

The tested biocide concentration achieved a substantial reduction (p < 0.05) in both total and free ATP values in the planktonic and biofilm samples after exposure to the biocide compared to that observed for the control (Figure 2(a)–(d)). No definitive differences in the tested surfaces were observed in terms of the reduction in the ATP concentration.

According to the DAPI/CTC fluorescence staining counts, after a 3 h biocidal treatment, the percentage of live bacteria was reduced to 1.50–2.63% within the first four months in the planktonic samples (Figure 3(a)). In contrast, after the fifth and sixth months, the live bacterial ratio increased to 7.41% and 8.51% of the total cells, respectively (Figures 3(b)–3(d)).

The results of supportive tests (ATP measurements and especially fluorescence staining) showed that the VBNC state was induced in biofilm-associated bacteria by biocidal treatment during biofilm maturation. While an increase in the log reduction for the cultivation results was observed, in contrast, the number of viable bacteria in the biofilm samples increased for all the tested materials. In addition, the absence of cultivable L. pneumophila after exposure to the same biocide concentration from the third month onward suggested that legionellae may have entered to the VBNC state. These results indicate that legionellae behaviour was more dependent upon biofilm density than on the surface material type, consistent with the results of Långmark et al. (25).

According to public health regulations, testing for legionellae is required at least once every three months, with acceptable limits reported as <10 CFU ml⁻¹ and <10⁴ CFU ml⁻¹ for legionellae and HPC, respectively. In addition, <300 RLU ATP values indicate clean tower conditions. The biocide tested in this study provided clean tower conditions for HPC, legionellae and ATP. The correlation between HPC and microbial ATP was low (p < 0.05, r = 0.570) and moderate between HPC and fluorescence staining counts (p < 0.05, r = 0.597). A high correlation was observed between ATP and fluorescence staining test results (p < 0.05, r = 0.986). Although there are no consensus values or standard limits for viable cell counts after fluorescence staining or for the risk assessment limit values of biofilm samples, for effective evaluations of the microbiological load of systems, free and total ATP measurement and fluorescence staining methods may be useful. Furthermore, according to the fluorescence staining results after biofilm maturation or the settlement of pollution in water systems, evaluations based on culture-based methods may be both time consuming and fail to reflect the actual conditions.

Since bioadhesion to surfaces and subsequent biofilm formation are of crucial importance with respect to both public health and industrial equipment fouling, evaluations of surface-bacteria interactions are essential. For this reason, three different substrata (PVC, glass and stainless steel) were monitored and compared for their responses to biocide exposure. Erdem et al. (20) showed that plastic polymers, such as PVC, polypropylene (PP) and polyethylene (PE) materials, supported the lowest HPC values when comparing glass and stainless steel in a model cooling tower system. Similarly, Hyde et al. (30) compared the adhesion rates of biofilms formed by a single species to different surfaces and observed the following order of adhesion (from lowest to highest): injection moulded PFA < polyvinylidene fluoride (PVDF) < rotationally moulded PFA < perfluoroalkoxy tetrafluoroetyhylene (PFA) < silicone < glass < polypropylene < stainless steel. Hyde et al. (30) also showed that the persistence of biofilms to removal by sodium hypochlorite exhibited the same order as the adhesion rate. Dos Santos et al. (31) observed that the abundance of sessile HPC was higher on stainless steel and glass surfaces than on carbon steel coupons in two industrial-scale cooling towers. However, the data in these studies were only obtained using cultivation-based methods. Furthermore, in the study by Hyde et al. (30), mixed species or natural biofilms were not evaluated.

In contrast to these results, Zacheus et al. (32) observed no significant differences in biofilm quantity on various surfaces (PVC, PE and stainless steel) with respect to HPC, acridine orange staining and organic carbon quantity results. In the current study, according to the observed Kruskal-Wallis asymptotic significance values, there was no significant difference between the sessile microbiota on the tested materials in terms of the HPC, fluorescence staining and total or free ATP level results (0.160, 0.080 and 0.298, respectively). The responses of biofilms to the oxidising biocide were similar to those described in previous reports, both for the microbial load on different surfaces and biofilm behaviour to the biocide (9, 32).

These results indicate that both the attachment and removal of biofilms may not be controlled by a single factor. Thus, several factors must be taken into consideration when evaluating a material for industrial use, including: (i) its effect on the intended and random microorganisms; (ii) the age and stage of the biofilm; (iii) the interactions between microorganisms and their ecological behaviour in the biofilm ecosystem; (iv) the mode of action of the applied biocide; (v) potential chemical interactions between the biological components and the applied biocide; and (vi) the surface finish and roughness of the substratum (16, 25, 30, 33). In these studies, differences observed between substrata were potentially based on the possible effect of the uptake of elements from the substratum by bacteria during biofilm adhesion or formation or with the corrosion of the materials through an increase in disinfectant demand by corrosion products (18, 19, 30, 32). Kröpfl et al. (19) made comparisons of biofilms grown on different substrata and found that biofilm production and abundance of algae were influenced by the substratum parallel with the concentrations of essential elements and heavy metal pollutants (zinc, nickel, lead and copper). On the other hand, the authors also emphasised that from a bacteriological point of view no substratum effects were detectable (19). Additionally, Zacheus et al. (32) have found that there was no difference in total cell count on PVC, PE and stainless steel surfaces before and after ozone treatment. In our study, from a bacteriological point of view no substratum type effects were detectable on the bacterial colony counts like Kröpfl et al.’s study (19) and response to the biocidal treatment, like Zacheus et al. (32).