Nobuyasu Yamaguchi, Graduate School of Pharmaceutical Sciences, Osaka University, 1–6, Yamada-oka, Suita, Osaka 565–0871, Japan (e-mail: firstname.lastname@example.org).
Physiologically active bacteria in purified water used in the manufacturing process of pharmaceutical products were enumerated in situ. Bacteria with growth potential were enumerated using the micro-colony technique and direct viable counting (DVC), followed by 24 h of incubation in 100-fold diluted SCDB (Soybean Casein Digest Broth) at 30 °C. Respiring and esterase-active bacteria were detected by fluorescent staining with 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) and 6-carboxyfluorescein diacetate (6CFDA), respectively. A large number of bacteria in purified water retained physiological activity, while most could not form colonies on conventional media. The techniques applied in this study enabled bacteria to be counted within 24 h so results could be available within one working day. These rapid and convenient techniques should be useful for the systematic monitoring of bacteria in water used for pharmaceutical manufacturing.
Several grades of water are used during pharmaceutical manufacturing, as a component of pharmaceutical products and for washing equipment. The grade of water should be selected according to its role in the process. Water must therefore meet strict quality standards, and that quality should be maintained until use.
The validation of manufacturing processes is particularly important to assure the quality of pharmaceutical products. Validating a manufacturing support system, including water supplies, may reduce dependence on intensive in-process testing and finished product testing. It is particularly critical to control the quality of the water used to prepare sterile medical products. In other words, manufacturing processes for water used to prepare sterile medical products should be of a high microbial quality. Therefore, the bioburden of the water must be rapidly and precisely determined. Conventional culture methods with incubation have often been employed to enumerate bacteria in pharmaceutical water by macroscopic observation. However, this method takes several days and is too slow for efficient quality control. To monitor on-line processes, a rapid method for enumerating bacteria is needed that provides results within a time frame that permits adjustment.
Recent studies describing the persistence of bacteria in aquatic environments have demonstrated that many of these organisms enter an altered physiological state termed the viable but non-culturable (VBNC) state ( Roszak & Colwell. 1987; Byrd et al. 1991 ; Yamaguchi et al. 1997 ). This is commonly defined as the inability to form colonies on media while retaining physiological activity. This is particularly important when assessing pathogenic bacteria in the VBNC state, which may be undetectable by standard cultivation on media but which remain viable ( Colwell et al. 1985 ; Rahman et al. 1996 ).
In this study, the physiological activity of bacteria was examined in purified water, one type of pharmaceutical grade water, both quantitatively and rapidly, as the first stage of microbiological monitoring during the pharmaceutical manufacturing process.
Materials and methods
The DNA-binding fluorochrome, 4′,6-diamidino-2-phenylindole (DAPI) and the esterase indicator fluorescent dye, 6-carboxyfluorescein diacetate (6CFDA), were purchased from Sigma. The redox dye, 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) was purchased from Polysciences Inc.
Water samples were collected in sterile polycarbonate bottles from the same point on the flow of the pharmaceutical water supply system and immediately analysed microbiologically.
Bacterial strains and growth conditions
Burkholderia cepacia ATCC25416T and Staphylococcus epidermidis IFO3762 were aerobically cultured in Soybean Casein Digest Broth (SCDB; 17 g peptone for casein, 3 g peptone for soy bean, 5 g sodium chloride, 2·5 g potassium hydrogen phosphate, 2·5 g glucose l−1 laboratory quality water) at 30 °C.
Total direct count and enumeration of colony-forming bacteria
Total bacteria (both live and dead cells) in purified water were counted by DAPI staining. Bacterial cells in purified water were vacuum-filtered onto a polycarbonate black filter (ADVANTEC, Tokyo, Japan; pore size 0·20 μm) in a funnel, then incubated in DAPI stock solution (10 μg ml−1 in sterilized water; final concentration; 1 μg ml−1) for 5 min at room temperature. The solution in the funnel was vacuum-filtered, then the filter was air-dried and mounted on glass microscope slides with non-fluorescence immersion oil (Olympus, Tokyo, Japan). The total number of bacteria was determined using an epifluorescence microscope (BH2; Olympus) equipped with a dichroic mirror (DM400), as well as excitation (UG1) and absorption (L420) filters.
The number of colony-forming bacteria was determined by filtration using the Milliflex-100 system (Millipore; pore size 0·45 μm). Bacteria trapped on filters were incubated at approximately 30 °C for 72 h on the filter paper soaked with SCDB. Bacterial colonies were counted macroscopically.
Enumeration of micro-colony-forming bacteria
Cells in purified water were vacuum-filtered onto a polycarbonate black filter (ADVANTEC; pore size 0·20 μm), which was immediately removed from the funnel and placed on the surface of filter paper (Whatman no. 2) soaked with 100-fold diluted sterilized SCDB. After a 24 h incubation at 30 °C, bacteria on filters were fixed on the surface of filter paper (Whatman no. 2) soaked with 4% (w/v) paraformaldehyde at 4 °C. The underside of the filter was gently washed with sterilized water, then placed uppermost in Petri dishes on the surface of 10 μg ml−1 DAPI containing polysorbate 80 (final concentration; 2% (w/v)) for 5 min (polysorbate 80 was added to enhance staining efficiency). Excess DAPI was removed by gently agitating the filter on a sterilized water surface to avoid colony disruption. Water droplets were air-dried by placing the filters in sterilized Petri dishes. Stained bacteria were enumerated by epifluorescence microscopy as well as by total direct counting.
Enumeration of viable bacteria with growth potential
The numbers of viable bacteria with growth potential were enumerated using the modified method of Kogure et al. (1978) . Cells in purified water were vacuum-filtered onto polycarbonate black filters (ADVANTEC; pore size 0·20 μm), which were immediately removed from the funnel and placed on filter papers (Whatman no. 2) soaked with a mixture of 100-fold diluted SCDB and nalidixic acid (NA; Sigma) in distilled water (final concentration; 20 μg ml−1). After 24 h of incubation at 30 °C, the filters were fixed on the surface of filter papers (Whatman no. 2) soaked with 4% (w/v) paraformaldehyde at 4 °C. Trapped bacteria were then stained and counted by epifluorescence microscopy as well as by micro-colony-forming bacteria.
Enumeration of respiring bacteria exhibiting CTC reduction
Bacterial cells were trapped by vacuum onto a polycarbonate black filter (ADVANTEC; pore size 0·20 μm) and the filter was immediately placed on the surface of filter paper (Whatman no. 2) soaked with SCDB (final 100-fold diluent) and CTC solution (final concentration; 10 mmol l−1). After 2 h of incubation at 30 °C in the dark, bacteria were stained with DAPI. Filters were then air-dried and mounted on glass microscope slides with non-fluorescing immersion oil (Olympus). Bacterial cells were enumerated using an epifluorescence microscope (BH2; Olympus) equipped with a 100 W mercury lamp. Respiring bacteria were viewed with an excitation filter (BP490 + EY455), a dichroic mirror (DM500) and an absorption filter (O515). DAPI-stained bacteria were observed as well as total direct counting. Thus, respiring and total bacterial cells were enumerated under blue light and u.v. excitation, respectively.
Enumeration of esterase-active bacteria by 6CFDA hydrolysis
Bacterial cells were trapped by vacuum onto a polycarbonate black filter (ADVANTEC: 0·20 μm) then incubated in 6CFDA buffer (0·1 mmol l−1 phosphate buffer, pH 8·5, 5% (w/v) NaCl, 0·5 mmol l−1 EDTA) ( Yamaguchi et al. 1997 ) and DAPI stock solution (10 μg ml−1 in sterilized water; final concentration 1 μg ml−1) in a filtration funnel for 5 min at room temperature. Thereafter, 6CFDA stock solution (10 mg ml−1 in acetone) was added to the filtration funnel to a final concentration of 150 μg ml−1 for 2 min at room temperature. The mixture in the funnel was then vacuum-filtered. Stained bacteria were counted by epifluorescence microscopy. Stained bacteria were enumerated under blue light and u.v. excitation by epifluorescence microscopy, as well as enumeration of CTC-reducing bacteria.
Substrate concentration for bacterial growth
Bacteria in purified water for the manufacture of pharmaceutical products were trapped onto a polycarbonate filter and their growth in different media investigated. Suitable media were assessed according to the number of micro-colonies. Table 1 shows the total bacterial number (TDC), the number of micro-colonies (m-cfu), and the percentage of micro-colonies to TDC after a 24 h incubation at 30 °C on various media; R2A medium ( Reasoner & Geldreich 1985), SCDB medium, 0·025% (w/v) yeast extract (YE), nutrient broth (NB) and plate count broth (PCB) at various dilutions were tested. In general, SCDB and PCB are widely used to enumerate various kinds of micro-organisms, and SCDB is routinely applied to the microbiological monitoring of pharmaceutical water; R2A medium and NB are frequently used to detect bacterial cells in the aquatic environment, and YE is infrequently used in direct viable counting to estimate bacterial viability. As shown in Table 1, R2A medium and 100-fold diluted SCDB were appropriate media for bacterial growth.
Table 1. Numbers of micro-colonies formed on various media
4·7 × 103
5·1 × 103
6·0 × 103
4·9 × 103
5·5 × 103
5·8 × 103
5·7 × 103
5·5 × 103
6·3 × 103
4·7 × 103
5·4 × 103
1·5 × 102
5·1 × 101
1·1 × 102
6·1 × 101
1·4 × 102
1·8 × 102
1·1 × 102
9·2 × 101
5·7 × 101
9·2 × 101
1·1 × 102
(Unit: cells ml−1). TDC, Total bacterial number; m-cfu, number of micro-colonies; SCDB, Soybean Casein Digest Broth; YE, 0·025% (w/v) Yeast Extract; NB, Nutrient Broth; PCB, Plate Count Broth.
Three times more micro-colonies formed on 100-fold diluted SCDB than stock SCDB; R2A and 100-fold diluted SCDB are recommended for enumerating viable cells in purified water, and 100-fold diluted SCDB was used in this study to enumerate bacteria in purified water. Dilute SCDB allowed small colonies to grow but prevented larger colonies from excessive growth through limiting diffusion of medium components.
Enumeration of micro-colony-forming bacteria and direct viable count
The optimal concentration of DAPI for enumerating bacteria in purified water was investigated. The number of bacteria stained with 1, 5, 8 or 10 μg ml−1 of DAPI for 5 min was determined ( Fig. 1). The total bacterial number reached a plateau when bacteria were stained with more than 5 μg ml−1 DAPI. However, DAPI fluorescence was brightest at 10 μg ml−1 DAPI. Bacteria were stained with 10 μg ml−1 DAPI in the following experiments.
The incubation period required for micro-colony formation and DVC in purified water was investigated up to 48 h. The highest number of micro-colonies obtained after 24 h of incubation and micro-colony was overgrowth after 36 h (data not shown). Bacteria subject to the DVC procedure were not elongated by a short incubation, and quinolone-resistant bacteria began to grow after 36–48 h. The ideal incubation period for routine microbiological monitoring needs to be defined, although suitable incubation periods were altered slightly in every sample. The number of micro-colonies and elongated (or fattened) bacteria was highest after a 24 h incubation, whereas the total bacterial number remained constant. Consequently, the optimal staining conditions for enumerating micro-colonies and DVC are: 10 μg ml−1 DAPI with 2% (w/v) polysorbate 80 (final concentration) for 5 min after a 24 h incubation.
CTC concentration and incubation period on the enumeration of respiring bacteria
The staining efficacy of CTC is affected by the structure of the cell wall and/or cell membrane, physiological state and species of target bacteria. The population in purified water may consist of various bacteria in different growth phases. We defined optimal conditions of CTC staining by Staph. epidermidis (Gram-positive) and B. cepacia (Gram-negative). Staphylococcus epidermidis is indigenous to human skin, and B. cepacia may contaminate distilled water ( Carson et al. 1973 ; Rapkin 1976). CTC reduction by bacteria at five growth phases (lag, early log, mid log, late log and stationary) was also examined. As shown in Fig. 2, less than 6% of Staph. epidermidis cells were stained at different phases. On the other hand, the ratio of CTC reducing to TDC depended on the concentration of CTC, and remained almost the same with a high concentration of CTC (10 mmol l−1) at all phases of B. cepacia.
In addition, the incubation period for optimal CTC staining was determined by B. cepacia (stationary phase) in 100-fold diluted SCDB (final concentration) containing 10 mmol l−1 CTC. The ratio of CTC reducing to TDC in various incubation periods ranging from 0·5 to 4 h are shown in Fig. 3(a). There was no significant difference between any of the incubation periods with respect to either the number of total bacteria as determined by DAPI staining, or the number of respiring bacteria. The incubation period was similarly investigated using six samples of purified water from the manufacturing process. As shown in Fig. 3(b), a 2 h incubation appeared to be sufficient for quantifying active cells, and longer incubation periods did not yield significantly higher numbers.
In view of the toxic effect of CTC on bacteria ( Ullrich et al. 1996 ), the optimal concentration of this redox dye, and the incubation period required for its reduction, were 10 mmol l−1 and 2 h, respectively.
6CFDA concentration and incubation period on the enumeration of esterase-active bacteria
Optimal 6CFDA staining conditions were determined using Staph. epidermidis and B. cepacia. Figure 4 shows the ratio of 6CFDA-stained bacteria to the TDC at various growth phases of Staph. epidermidis and B. cepacia. Although the number of 6CFDA-stained Staph. epidermidis cells was slightly decreased at mid log and late log phase, at all phases, most cells were stained by 6CFDA. On the other hand, the ratio of 6CFDA-stained to total B. cepacia cells depended on the concentration of 6CFDA up to 150 μg ml−1; at higher concentrations, the ratio of 6CFDA-stained bacteria to TDC remained almost the same or decreased slightly at all phases except early log phase.
The effect of the incubation period on B. cepacia and purified water samples from the manufacturing process in 150 μg ml−1 of 6CFDA was also examined ( Fig. 5). The incubation period had no significant effect on either the number of total bacteria as determined by DAPI staining, or the number of esterase-active bacteria in B. cepacia ( Fig. 5(a)). However, the ratio of 6CFDA-stained bacteria to TDC in purified water samples was reduced after an extended staining period ( Fig. 5(b)) because the green fluorescent background in the microscopic fields gradually increased. The optimal 6CFDA concentration and incubation period for 6CFDA staining were 150 μg ml−1 and 2 min, respectively.
Bacterial numbers in purified water used in the pharmaceutical manufacturing process
The number of bacteria in purified water used in the pharmaceutical manufacturing process was determined by total direct counting, the conventional culture technique, direct viable counting, the micro-colony technique and vital staining ( Fig. 6). Purified water samples were collected from October to November 1997. There were no significant differences in the bacterial number of eight purified water samples. Total bacterial numbers were 3·9–9·4 × 105 cells 100 ml−1 and numbers of colony-forming bacteria were 5–260 cfu 100 ml−1. DAPI staining detected 103–105-fold more total bacteria than the conventional colony-forming method. The numbers of esterase-active bacteria were 1·4–5·1 × 103 cells 100 ml−1. Approximately half the total bacterial population retained esterase activity. Bacteria with respiratory activity, micro-colonies and viable bacteria with growth potential represented 0·5–11, 1–5 and 4–17% of the TDC, respectively.
Nutrient limitation, desiccation ( Pedersen & Jacobsen 1993) or metabolite accumulation may induce an adaptive physiological stress response in bacteria, and many organisms adapt an altered physiological state in an aquatic environment and are only viable under nutrient-poor conditions ( Byrd et al. 1991 ). Nutrient-rich conditions may inhibit the growth of these bacteria, while purified water contains insufficient nutrient for bacterial growth. Thus, 100-fold diluted SCDB, which is composed of low levels of organic substances and various minerals, was the most suitable medium for micro-colony formation. We determined the optimal CTC and 6CFDA staining concentrations using Staph. epidermidis (Gram-positive) and B. cepacia (Gram-negative) at five growth phases; CTC and 6CFDA staining efficacy was affected by bacterial species and growth phase, due to differences in permeability to CTC and 6CFDA.
Bacterial cells (105–106), many of which could not be detected by the membrane filter method followed by culture, were found in 100 ml purified water used in the manufacturing process. These bacteria exhibited physiological activity such as respiration, enzyme activity and growth potential. One reason for the difference between the bacterial number determined by the membrane filter method and by fluorescent staining is that the majority of cells are in the VBNC state under nutrient-poor conditions, such as in purified water. The fact that most bacteria in water are under starvation conditions is widely accepted. Many of these bacteria remain metabolically active but are unable to grow on established media, i.e. they are in the VBNC state. Therefore, bacteria in purified water consist of (i) culturable cells that can form colonies on media, (ii) viable but non-culturable cells which are metabolically capable but which cannot form colonies and (iii) dormant or dead cells that are visible by microscopy but have no metabolic activity ( Pedersen & Jacobsen 1993). As culturability may also depend on growth medium, the distinctions of bacterial state described above are difficult. Moreover, few studies have attempted to examine the physiological activity of bacteria in the VBNC state. Further study is required to understand bacteria in the VBNC state. However, bacteria with physiological activity such as growth potential and enzymatic activity can have major effects on pharmaceutical products and should be accurately detected.
The in situ enumeration methods described in this study can be used to detect bacteria in the VBNC state and are therefore recommended for bioburden assessment. In this study, three physiological activities (growth potential, respiration and enzymatic activity) were noted, and four enumeration techniques (micro-colony technique, direct viable counting, CTC- and 6CFDA-staining) were examined to determine the bacterial number in purified water. Direct viable counting has several disadvantages associated with the existence of bacteria resistant to nalidixic acid ( Buchrieser & Kaspar 1993; Joux & Le Baron 1997 ), the optimal concentration of nalidixic acid, and the objective estimation of bacterial elongation. On the other hand, when using the micro-colony technique, filters may become overgrown with rapidly growing bacteria if the sample contains a large number of culturable cells. However, bacteria with growth potential can be detected within 24 h using the micro-colony technique and DVC method, and these methods enable us to detect bacterial cells in the VBNC state. Even though staining efficacy of CTC differs according to the bacterial species or growth phase, respiring bacteria can be detected within only 2–3 h by CTC-staining. Moreover, detection of bacteria with esterase activity takes only 30 min by 6CFDA-staining. These techniques are therefore useful for routine enumeration of bacteria in the purified water used in pharmaceutical manufacturing processes. Data analysis can be expedited and results are available within one working day, which should allow for better control of water supplies where contamination is a concern.
In addition, filtration of a larger amount of water, or use of a scanning cytometer, would improve on the current detection limit of 103 cells ml−1. The scanning cytometer can also enumerate bacteria in other grades of water used in pharmaceutical manufacturing such as reverse osmosis, distilled and ultra-filtration water. Thus, these rapid and convenient techniques should be useful for the systematic bacterial monitoring of water used in the pharmaceutical industry.