Abstract

Microorganism growth in industrial systems is controlled through the use of biocides and biodispersants. There is, however, no simple means of determining the efficacy of these control mechanisms, but it is currently tested using complex bacterial culturing techniques. Biolog Ecoplates® have been used to detect bacterial population changes in various communities. These microtitre plates comprise 31 different carbon substrates (in triplicate) with wells. When a sample is added to the wells, bacteria capable of metabolising the relevant carbon sources respire the substrates, causing the tetrazolium dye in the well to turn purple, indicating a positive result. Hypothetically, the higher the microbial diversity, the more substrates will be utilised and vice versa. The objective of this study was to test this hypothesis, using Biolog Ecoplates® as a potential simple indicator to determine the efficiency of a biocide to control microbial growth in cooling water systems by monitoring the changes in the microbial metabolic pattern. This study proved the hypothesis using Biolog Ecoplates®, indicating that the addition of biocides at various concentrations resulted in fewer substrates being utilised, indicative of a decrease in microbial species diversity.

INTRODUCTION

South African power plants are mostly wet cooled, obtaining their water from various surface water sources including dams and rivers (Kotze 2015). Due to the poor water quality of these surface waters, chemical treatment of the cooling water is required to prevent biofouling and corrosion. Treatment includes the dosing of biodispersants and biocides to minimise and control microbial growth and biofouling within the cooling water system (Satpathy et al. 2016).

Bacteria, like all organisms, are affected by stress in their environment (Martinez 2013). It is theorised that the addition of a biocide (a source of stress) to the environment should decrease the number of species present (the species diversity) and numbers of bacteria (the abundance). Dosage concentration, the type of chemical dosed and contact time has an effect on this decrease (Satpathy et al. 2016).

Although several biocides are used by the South African power generation industry, isothiazalone and dibromonitrilopropionamide (DBNPA) are the most common. However, bacteria may become resistant to isothiazolone (Zhou et al. 2016). Hence, these biocides are usually applied in an alternating pattern to avoid possible development of bacterial resistance (Eskom Legionella Standard 2017). DBNPA is a fast acting, non-oxidising (despite the bromine) biocide that inactivates enzymes, destroying metabolic pathways responsible for energy production and transfer of wastes and nutrients. It is neutral and degrades to relatively innocuous by-products (Williams & McGinley 2010). Due to its lack of persistence it is most often used as an alternative biocide. It becomes deactivated above pH 8.5 or in the presence of sulphide or sulphite contamination (Huber et al. 2010).

Historically at Eskom, biocide dosing was conducted on a two-weekly cycle with a continuous biodispersant dosage. Biocide efficacy testing was done by conducting classical plate counts on the planktonic bacterial population in cooling water (Eskom Standard). However, many bacterial species are non-culturable by classical microbiology methods, thus an inaccurate indication of the bacterial population in the treated water is likely (Parshionikar et al. 2009). In this study, two experimental methods have been employed that do not rely on artificial culturing techniques (in a Petri dish on a solid medium) and instead focus on implying species/population diversity by metabolic and genetic means.

By way of a conventional metabolic-based microbiology approach, Biolog Ecoplates® have been successfully used to monitor changing populations in water (Lekhanya 2010), sewage (Gryta et al. 2014) and soil (Xu et al. 2015). Biolog Ecoplates® contain a triplicate set of 31 carbon substrates, with a blank well in each set. When the bacteria utilise the substrate in the well they respire and cause a colour change in the dye present in the well.

The premise of this as a method to determine microbial species diversity within different samples relies on the organisms' ability to utilise various carbon substrates/sources for nutrition. If bacterial species, in a certain environment, are killed or altered through the addition of a stress to the environment (in this case a biocide), the number of substrates utilised should decrease (Zhou et al. 2016).

Cellular genomic-based methods may be employed in the detection of bacterial diversity and potentially the full description of complex microbial communities. Among these techniques is denaturing gradient gel electrophoresis (DGGE), which uses the genetic sequences of organisms in a target environment in order to detect microbial community diversity (Mayrhofer et al. 2014). Genetic variations within microorganisms' genes and possible mobility shifts cause differentiation of genetic sequences, allowing species in environmental samples to be separated (Carmona et al. 2012; Shah 2015). Based on the sensitivity of double stranded DNA (dsDNA) to heat and chemicals, target populations within a sample can be separated into a visual genetic fingerprint. This method thus allows for a highly sensitive means of determining community diversity.

Species diversity is not necessarily linked to abundance in an environment. For example, high numbers of bacteria normally occur in extreme environments, although the species diversity is normally very low. It is therefore obvious that bacterial species diversity is a better indicator of stress in an environment than the total count of bacteria in the environment.

Hypothetically, the higher the bacterial diversity, the more substrates will be utilised and a lower diversity will lead to fewer substrates being utilised. An effective biocide will lead to the reduction of the microbial species diversity and abundance and hence fewer substrates utilised. The objective of this study was to test this hypothesis, using Biolog Ecoplates® as a potential simple indicator to determine the efficiency of a biocide to control microbial growth in cooling water systems by monitoring the changes in the microbial ecology.

MATERIALS AND METHODS

A 500 L cooling water system simulator (CWSS) (Figure 1) was used to evaluate changes in bacterial species diversity after the addition of DBNPA biocide at concentrations of 8 mg/L and 20 mg/L. The CWSS included a storage drum (1), a circulating pump (5), a heating vessel (4), which heats the water to 42 °C; an ICT 850 cooling tower supplied by Industrial Cooling Towers, Alrode, Johannesburg (6) and a pressure stabilisation drum (3), for water level maintenance. A stopcock (2) was used to ensure that any water losses incurred, either by evaporation or windage, were corrected. The CWSS could be drained through a drain valve to a collection drum (7) if the water in the system needed to be replaced.

Figure 1

Flow diagram of the cooling water simulator system (CWSS).

Figure 1

Flow diagram of the cooling water simulator system (CWSS).

The CWSS was filled with cooling water from an operational power plant and the water was circulated, without treatment, for 3 days prior to any biocide dosage. DBNPA biocide was added after taking an initial pre-dose sample (time 0). Samples were then taken after 15 and 30 min, and 1, 2, 3, 4, 6, 24 and 48 hours post dosing. The classical microbiology and Biolog Ecoplate® analyses were done by the Eskom Research, Testing and Demonstration microbiology laboratory and the molecular work done by MicroSci Consulting.

Classical microbiology

Total aerobic bacteria plate counts were conducted by completing 10-fold serial dilutions in sterile Ringers solution; 1 mL of each dilution was aseptically transferred into a sterile 90 mm Petri dish. Approximately 15 mL of cooled, molten plate count agar (Oxoid Ltd, Basingstoke, Hampshire, UK) was added to the Petri dish after which the Petri dish was swirled gently to mix the sample into the agar and allowed to set. The Petri dishes were incubated, inverted, at 37 °C for 48 hours. This method was utilised because it is the methodology most commonly employed on the power plants to determine total bacterial counts. All colonies that formed on the agar were counted, and this count was multiplied by the dilution factor to determine the final count.

Substrate utilisation – Biolog ecoplates®

Each of the 96 wells on the Biolog Ecoplate® (Biolog Inc., Hayward, CA, USA) was filled with 150 μL of sample. The Biolog Ecoplates® were then incubated at 35 °C and examined after 24 and 48 hours respectively. Any purple colouration, regardless of intensity, was marked and counted as a positive reaction. These reactions were logged (binary) and used to construct a digital graphic image of the plates for statistical analysis.

Molecular analysis

For molecular analysis, 300 mL of sample was filtered through a 0.45 μm sterile filter and then aseptically transferred into 20 mL of sterile saline solution.

Total DNA was extracted using the BIO101 Fast DNA Spin kit (Soil)® (Qbiogene Molecular Biology Products, Pretoria, South Africa).

A 1 g sample was added to Lysing Matrix E tubes thereafter, sodium phosphate buffer (978 μl) and MT buffer (122 μl) were added, the tube was vortexed for 30 seconds and centrifuged at 20,817 g (10,000 rpm) for 10 min.

The supernatant was transferred to a clean tube, 250 μl PPS reagent was added and mixed by inversion. The suspension was then centrifuged for 5 min at 20,817 g (10,000 rpm) to pellet the precipitate. The supernatant was transferred to a clean tube and 1 mL of binding matrix suspension added.

After settling, approximately 500 μL of supernatant was removed and discarded. The binding matrix was then resuspended in the remaining supernatant and 600 μL of the suspension was added to a SPIN™ Filter and centrifuged at 20,817 g (10,000 rpm) for 1 min. The catch tube was emptied and the remaining supernatant added to the SPIN™ Filter and centrifuged.

Subsequently 5 μL of SEWS-M was added to the SPIN™ Filter and centrifuged 20,817 g (10,000 rpm) for 1 min. The flow-through was decanted and the SPIN™ Filter replaced in the catch tube. The pellet was centrifuged at 20,817 g (10,000 rpm) for 2 min to dry the matrix.

The SPIN™ Filter was placed in a fresh catch tube and air dried for 5 min at room temperature. A total of 50 μL DES (DNase/pyrogen free water) was added to the matrix and gently stirred with a pipette tip. The resuspension was centrifuged at 20,817 g (10,000 rpm) for 1 min to transfer the eluted DNA to the catch tube.

Polymerase chain reaction (PCR) amplification

A portion of the bacterial 16S rRNA gene was amplified by PCR using K and M primers.

K (PRUN518R): 5′ATT-ACC-GCG-GCT-GCT-GG3′ (Surridge 2007)

M (pA8f-GC): 5′CGC-CCG-CCG-CGC-GCG-GCG-GGC-GGG-GCG-GGG-GCA-CGG-GGG-GAG-AGT-TTG-ATC-CTG-GCT-CAG3′ (Surridge 2007).

A reaction with no template DNA was included as a negative control. Each PCR tube contained a total volume of 20 μL: 10.8 μL sterile SABAX water, 2.5 μL PCR buffer (10×), 2 μL MgCl2 (25 mM), 2 μL dNTPs (2.5 μM), 1 μL primer K (50 μM), 1 μL primer M (50 μM), 1 μL template DNA (27 ng/μL), 0.2 μl Taq polymerase (5 U/μL). Prokaryotic DNA amplification was performed in a PCR thermal cycler using the following programme: 10 min at 95 °C, 35 cycles of 30 seconds at 94 °C, 30 seconds at 58 °C and 1 min at 72 °C, followed by 10 min at 72 °C, and then held at 4 °C. PCR product was analysed on a 1.5% TAE agarose gel.

Denaturing gradient gel electrophoresis

PCR products were subjected to DGGE according to the method described by Carmona et al. (2012). In short, 10 μl (ca. 250 ng) of each PCR product was loaded per lane onto a 25–55% urea/formamide denaturing gradient gel. Gels were run at 70 V for 17 hours at a constant temperature of 60 °C. Image analysis was performed using the Gel2 K programme (Norland 2004; Surridge 2007). The fingerprint comparisons were analysed using CLUST (Norland 2004; Surridge 2007). Dominant bands were compared and analysed to determine population diversity. Several comparison options are available but Jaccard and Simple indices (matchings) were used in this study.

These comparisons differ in how they compare the results obtained. Simple matching involves the comparison of the number (diversity) of species within a sample (DGGE lane) in comparison to other samples (lanes). Jaccard matching (Jaccard index/similarity coefficient) is a statistical comparison of similarity and diversity of sample sets. It is defined as the size of the intersection divided by the size of the union of the sample sets (Jaccard 1901).

RESULTS AND DISCUSSION

Substrate utilisation and classical microbiology

After dosing 20 mg/L of DBNPA, the total planktonic bacterial numbers decreased from 5.6 × 105 CFU/mL to 1.06 × 102 CFU/mL after 15 min (Figure 2). After dosing 8 mg/L of DBNPA the total planktonic bacterial numbers decreased from 1.74 × 105 CFU/mL to 2.3 × 102 CFU/mL after 1 hour (Figure 3). This indicated that the higher biocide concentration (20 mg/L) resulted in a shorter time period (15 min) to reduce the bacterial numbers than at the lower concentration (8 mg/L) where the same reduction was achieved only after 1 hour (Figure 3). The higher dosage also maintained the low counts for a longer period and slowed the relative regrowth over the 48 hour trial (Figure 2). These results confirm previous studies indicating the relationship between biocide concentration, contact time and efficiency (Netshidaulu 2015). An increase in biocide concentration normally results in a more rapid killing of microorganisms in a system (Williams 2007). Nevertheless, there is not necessarily a linear relationship between biocide concentration and kill rate (Wales & Davies 2015). The fact that the higher biocide concentration had a longer-lasting effect than the lower concentration was due to the extended half-life of the biocide at the higher concentration. (Taylor Industrial 2016). These results indicate the importance of dosing the correct concentration of a given biocide to control microorganisms in water systems and especially cooling water systems that are by design dynamic systems. Calculating the theoretical dosage level of a given biocide taking into consideration the minimum inhibitory concentration of the biocide is easy when based on the hydrodynamics of a system (i.e. system volume, dilution rate due to system losses, make-up water added, etc.) (Kanga 2010). However the efficiency of the biocide is also affected by non-hydraulic factors such as pH, interaction with organic matter in the system and overall system chemistry (Gnanadhas et al. 2012). Measuring the bioactivity of a biocide (the availability of the biocide to kill microorganisms in a particular system) is therefore more useful than doing only a theoretical hydrodynamic calculation, since it takes all system parameters into account (Tidwell & Broussard 2014). This was the reason why the carbon substrate utilisation approach was followed as part of this study.

Figure 2

Biolog Ecoplate® substrate utilisation versus aerobic bacterial counts for the duration of the 20 mg/L DBNPA trial.

Figure 2

Biolog Ecoplate® substrate utilisation versus aerobic bacterial counts for the duration of the 20 mg/L DBNPA trial.

Figure 3

Biolog Ecoplate® substrate utilisation versus aerobic bacterial counts for the duration of the 8 mg/L DBNPA trial.

Figure 3

Biolog Ecoplate® substrate utilisation versus aerobic bacterial counts for the duration of the 8 mg/L DBNPA trial.

In terms of the Huston (2014) species diversity model, these results indicate that the bacterial population with the 8 mg/L DBNPA dosage moves from a Category B environment (high bacterial numbers and high species diversity) to a Category A environment (low bacterial numbers and high species diversity), indicating an inadequate dosage concentration. The bacterial population from the 20 mg/L trial showed a bacterial population shift from Category B to Category C (low bacterial numbers and low species diversity), showing an effective biocide dose.

The reduction in carbon substrate utilisation at 15 min, 30 min and 1 hour after the 20 mg/L dose (Figure 4), is indicative that the bacterial species capable of utilising the substrates were metabolically inhibited. There were, however, still a small number of surviving bacteria present in the system at these times not reflecting as metabolically active based on the metabolic analysis (Figure 2). The reason for this is speculated to be due to the reduced bacterial numbers being below the detection level (sensitivity) of the Biolog Ecoplate® system. Unfortunately the exact detection level of the Biolog Ecoplate® has not been established. This may be due to bacteria remaining in the treated water not being able to utilise/metabolise the carbon substrates available.

Figure 4

Gel2 K statistical relationship of substrate utilisation at various times throughout the 20 mg/L DBNPA trial.

Figure 4

Gel2 K statistical relationship of substrate utilisation at various times throughout the 20 mg/L DBNPA trial.

After 2 hours the surviving bacteria again began to multiply, showing an increase in the total aerobe count and increased substrate utilisation to the original levels (Figure 2). These results are in agreement with previous studies in cooling towers (Brözel & Cloete 1992; Liu et al. 2011). Brözel & Cloete (1992) indicated that the bacterial numbers in cooling towers after regrowth, as in this study, could exceed the bacteria number before biocide dosage. The reason for this was given as the survival of a more resistant species of bacteria than was the survival of by the particular biocide used (Brözel & Cloete 1992; Chien et al. 2013).

Carbon substrate utilisation results for the 8 mg/L trial indicated the presence of metabolic diversity throughout the trial period after biocide dosage (Figure 4). This suggests that, although the DBNPA was effective at reducing the planktonic counts, there were still a variety of bacterial species that survived capable of utilising a variety of the various carbon sources. Similarly results were observed by Du Toit (2007) in paper effluent treated by DBNPA. This indicated that the biocide was ineffective at 8 mg/L in reducing the species diversity in the system. This was supported by the more rapid recovery of aerobic planktonic bacterial counts in the system compared to the higher dosage (Figure 3).

In order to generate a statistical comparison amongst the substrates utilised over the experimental period, simulated gel diagrams were developed for analysis with Gel2 K.

Substrate utilisation patterns for the 20 mg/L trial indicate a rapid decline in substrate utilisation between time 0 and 15 min after biocide addition, indicating the efficiency of the biocide (Figure 4). However, after 2 hours some substrate utilisation is evident increasing through 4 hours, 6 hours, 24 hours and eventually after 48 hours (Figure 4). This substantiates the recovery of microbial populations in the community. Substrate utilisation patterns for time 0 and after 48 hours showed only a 30% difference and clustered together distinctly different from the other time periods during the 48 hour trial (Figure 4).

This might account for the 30% difference that still existed in terms of substrate utilisation between time 0 and 48 hours. The results nevertheless indicate that the bacterial community was busy recovering to the original microbial diversity and population composition within the community at the start of the trial. It was concluded that the biocide did not eliminate the bacteria species present, but merely reduced the numbers to below the detection threshold of the Biolog Ecoplate® system. This result was also reported by Gryta et al. (2014), in an evaluation of the treatment of dairy sewage sludge.

In the 8 mg/L trial substrate utilisation never differed more than 30% during the different sampling times over the trial period, with the exception of the sample taken after 48 hours (Figure 5). The sample taken after 48 hours displays the lowest substrate utilisation and a 35% variation from the other samples, implying a species diversity change and decrease, plausibly either as a result of biocide efficacy or the stabilisation of the bacterial population once the dominant bacteria re-establish in the system, as proposed by Forbes et al. (2017). This indicates that the biocide at a concentration of 8 mg/L had a limited effect on the metabolic diversity and hence the species diversity in the system and that bacterial population changes in the community were marginal (Figure 5).

Figure 5

Gel2 K statistical relationship of substrate utilisation at various times throughout the 8 mg/L DBNPA trial.

Figure 5

Gel2 K statistical relationship of substrate utilisation at various times throughout the 8 mg/L DBNPA trial.

Molecular analyses results

Gel2 K was used to generate a DGGE gel pattern for statistical analysis of the 20 mg/L trial. The DGGE results indicate that there was no complete elimination of all of the bacterial species after treatment at 20 mg/L of DBNPA over the trial period (Figure 6). The DGGE results indicate that the bacteria diversity in the system was retained throughout the trial after biocide addition. This is in contrast with the Biolog Ecoplate® results that showed a complete inhibition of metabolic activity within 15 min of biocide dosage (Figure 2). The DGGE analysis supports the previous conclusion, that the biocide reduced bacterial numbers to below the detectable threshold required for substrate utilisation of the Biolog Ecoplate® system, but not the species diversity as indicated by the genetic diversity remaining in the system. This result is supported by work completed by Gryta et al. (2014). This explains the recovery of bacterial populations in the system and increased metabolic activity during the recovery stage within the 48 hour trial period (Figure 2). These results suggest that a viable but non-culturable state was induced at the lower biocide concentration. DGGE analysis hence may be a useful tool to indicate a viable but non-culturable (VBNC) state in bacteria, but will require further investigation.

Figure 6

The Gel2 K statistical relationship between the DGGE banding patterns at the various times throughout the 20 mg/L DBNPA trial.

Figure 6

The Gel2 K statistical relationship between the DGGE banding patterns at the various times throughout the 20 mg/L DBNPA trial.

As with the previous analyses, Gel2 K was used to generate a DGGE gel fingerprint pattern for statistical analysis of the 8 mg/L trial. These DGGE results indicate the presence of a diverse group of bacteria throughout the trial period after treatment with 8 mg/L of DBNPA (Figure 7). This is in agreement with the Biolog Ecoplate® results that showed continued metabolic activity after biocide treatment (Figure 3). This explains the recovery of bacterial populations in the system and increased metabolic activity during the recovery stage within the 48 hour trial period (Figure 3). These results also imply that the 8 mg/L biocide concentration was sub-optimal for use as a treatment regime and therefore ineffective, requiring a higher dosage concentration.

Figure 7

The Gel2 K statistical relationship between the DGGE banding patterns at the various times throughout the 8 mg/L DBNPA trial.

Figure 7

The Gel2 K statistical relationship between the DGGE banding patterns at the various times throughout the 8 mg/L DBNPA trial.

CONCLUSIONS

The addition of 20 mg/L of DBNPA resulted in a decrease in the number of bacteria within 15 min and reduced the metabolic activity of the bacteria to below the threshold of the Biolog Ecoplate®. DGGE analysis indicated that the species diversity was not reduced to the same extent as the bacterial numbers, resulting in a recovery of the metabolic diversity with the increase in bacterial numbers. The microbial ecosystem dynamics therefore shifted from a scenario of high species diversity and a high number of individuals within each species to a scenario of high species diversity and low numbers of individuals within each species. This does not reflect an ideal outcome for an effective biocide program (in this case DBNPA used at 20 mg/L). An ideal biocide, used at the optimal concentration and contact time, would have resulted in a reduction of both the number of species and the number of individuals within a species. This result was exacerbated when a lower concentration of the biocide (8 mg/L) was used, where only the number of individuals within each species was reduced but not the diversity of species, as indicated by both the Biolog Ecoplate® system and DGGE results.

In terms of the hypothesis that the Biolog Ecoplate® system could potentially be used as a simple method for determining the efficiency of a biocide, the results indicated that it could indeed be used with some caution when interpreting the results. The significant decrease in metabolic activity using the Biolog Ecoplate® system did not reveal the fact that many of the species were still present in the system after biocide addition, but at levels below the sensitivity of the system as supported by the DGGE results. However, the recovery of the bacterial numbers and metabolic diversity within 48 hours did confirm the latter. In addition, the DGGE results may indicate that the addition of a biocide induces a VBNC state within the bacteria and suggests that DGGE may be a useful tool to detect and monitor bacteria in a VBNC state.

This study indicated that monitoring of the metabolic activity (as an indirect indicator of species diversity), using a system such as the Biolog Ecoplate® in a water cooling tower ecosystem, is a more useful technique for monitoring the efficacy of a biocide than determining the total number of bacteria in a system, since the recovery of the metabolic diversity can be directly linked to the resilience of a system against a natural (extreme environment) and/or anthropogenic stress such as adding a biocide.

This means of determining biocide efficacy can therefore only be utilised when the biocide is being dosed at optimal levels for the system being tested. It is therefore essential that the supplier's recommendations be correctly followed.

The effect of other biocides used in cooling water treatment plants must be evaluated, and the accuracy tested on an operational plant. This will determine which other external factors may affect the biocidal efficacy.

REFERENCES

REFERENCES
Brözel
V.
&
Cloete
T. E.
1992
The effect of bactericide treatment on planktonic bacterial communities in water cooling systems
.
Water SA
18
(
20
),
87
92
.
Du Toit
R.-M.
2007
The Identification of Prevalent Bacterial Isolates and Characterisation of Microbial Communities in Paper-Mill Water Systems
.
MSc thesis
,
Department of Microbiology and Plant Pathology, Faculty of Natural and Agricultural Sciences, University of Pretoria
.
Eskom Legionella Standard
2017
Legionella Control and Management in Water Systems. Eskom Standard, 240-123919938
.
Forbes
S.
,
Cowley
N.
,
Humphreys
G.
,
Mistry
H.
,
Amézquita
A.
&
McBain
A. J.
2017
Formulation of biocides increases antimicrobial potency and mitigates the enrichment of nonsusceptible bacteria in multispecies biofilms
.
Applied and Environmental Microbiology
83
(
7
),
e03054-16
Gnanadhas
D. P.
,
Marathe
S.
&
Chakravortty
D.
2012
Biocides –resistance, cross-resistance mechanisms and assessment
.
Expert Opinion on Investigational Drugs
22
.
doi:10.1517/13543784.2013.748035
.
Gryta
A.
,
Frac
M.
&
Oszust
K.
2014
The application of the biolog ecoPlate approach in ecotoxicological evaluation of dairy sewage sludge
.
Applied Biochemistry and Biotechnology
.
doi:10.1007/s12010-014-1131-8
.
Huber
P.
,
Carre
B.
&
Zeno
E.
2010
Effects of biocides on wet-end chemistry
.
Bioresources
5
(
3
),
1675
1688
.
Jaccard
P.
1901
Étude comparative de la distribution florale dans une portion des Alpes et des Jura
.
Bulletin de la Société Vaudoise des Sciences Naturelles
37
,
547
579
.
Kanga
Y.
2010
Controlled Release of Isothiazoline Biocides From Industrial Minerals
.
PhD thesis
,
School of Chemical Engineering, College of Engineering University of Birmingham
.
Kotze
P.
2015
Water and energy: how much water will we need to keep the lights on?
The Water Wheel
14
(
3
),
22
25
.
Lekhanya
L. L.
2010
Bacterial Diversity of Soil Irrigated with Gypsiferous Mine Water as Determined by Culture-Dependent and Independent Techniques
.
MSc thesis
,
University of Pretoria
.
Liu
Y.
,
Zhang
W.
,
Sileika
T.
,
Warta
R.
,
Cianciotto
N. P.
&
Packman
A. I.
2011
Disinfection of bacterial biofilms in pilot-scale cooling tower systems
Biofouling
27
(
4
),
393
402
.
Martinez
J. L.
2013
Bacterial pathogens: from natural ecosystems to human hosts
.
Environmental Microbiology
15
(
2
),
325
333
.
Mayrhofer
S.
,
Filipp
P.
,
Lehner
D.
,
Reiterich
C.
,
Kneifel
W.
&
Domig
K. J.
2014
Suitability of different PCR-DGGE primer sets for the monitoring of lactic acid bacteria in wine
.
South African Journal for Enology and Viticulture
35
(
2
),
185
195
.
Netshidaulu
A. E.
2015
Impact of Chlorine and Wastewater Contact Time, Chlorine Residual and Mixing on Microorganism Inactivation
.
MSc Civil Engineering, University of the Witwatersrand, Johannesburg, South Africa
.
Norland
S.
2004
Gel2 K gel Analysis Software
.
University of Bergen
,
Norway
. .
Parshionikar
S.
,
Haugland
R. A.
,
Shanks
O.
,
Hunt
M. E.
,
Cottrill
M.
,
Grimm
A.
,
Genthner
O. H. F.
,
Harris
S.
,
Oshiro
R. K.
,
Lincoff
A.
,
Pfaller
S.
&
Sivaganesan
M.
2009
Method Validation of U.S. Environmental Protection Agency Microbiological Methods of Analysis
.
Prepared for: The EPA Forum on Environmental Measurements (FEM) by The FEM Microbiology Action Team, FEM Document Number 2009–01
.
Satpathy
S.
,
Kumar Sen
S.
,
Pattanaik
S.
&
Raut
S.
2016
Review on bacterial biofilm: an universal cause of contamination
.
Biocatalysis and Agricultural Biotechnology
7
,
56
66
.
Shah
J.
2015
Exploitation of denaturing gradient gel electrophoresis in analysis of microbial diversity
.
Journal of Microbial & Biochemical Technology
7
,
112
119
.
Surridge
A. K. J.
2007
Denaturing Gradient Gel Electrophoresis Characterisation of Microbial Communities in Polycyclic Aromatic Hydrocarbon and Polychlorinated Biphenyl Contaminated Soil
.
PhD Thesis
,
University of Pretoria
,
South Africa
.
Taylor Industrial
2016
Calculating HTI Allows You to Properly Protect and Operate Open Cooling Tower Systems
. .
Tidwell
T. J.
&
Broussard
Z. R.
2014
Methods of determining biocide efficacy or mechanism of action using flow cytometry. US Patent Office, US 9382572 B2
.
Williams
T. M.
2007
The mechanism of action for isothiazalone biocides
.
Powerplant Chemistry
9
(
1
),
14
22
.
Williams
T. M.
&
McGinley
H. R.
2010
Deactivation of industrial water treatment biocides
.
CORROSION 2010
,
14–18 March
,
San Antonio
.
Texas NACE International
.
Xu
W.
,
Ge
Z.
&
Ram Poudel
D.
2015
Application and optimization of biolog EcoPlates in functional diversity studies of soil microbial communities
.
MATEC Web of Conferences. 10.1051/matecconf/20152204015
.
Zhou
G.
,
Shi
Q.-S.
,
Huang
X.-M.
&
Xie
X.-B.
2016
Comparison of transcriptomes of wild-type and isothiazolone-resistant Pseudomonas aeruginosa by using RNA-seq
.
Molecular Biology Reports
43
.
doi:10.1007/s11033-016-3978-y
.