• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Apr 2002; 68(4): 1548–1555.
PMCID: PMC123900

Analysis of Bacteria Contaminating Ultrapure Water in Industrial Systems


Bacterial populations inhabiting ultrapure water (UPW) systems were investigated. The analyzed UPW systems included pilot scale, bench scale, and full size UPW plants employed in the semiconductor and other industries. Bacteria present in the polishing loop of the UPW systems were enumerated by both plate counts and epifluorescence microscopy. Assessment of bacterial presence in UPW by epifluorescence microscopy (cyanotolyl tetrazolium chloride [CTC] and DAPI [4′,6′-diamidino-2-phenylindole] staining) showed significantly higher numbers (10 to 100 times more bacterial cells were detected) than that determined by plate counts. A considerable proportion of the bacteria present in UPW (50 to 90%) were cells that did not give a positive signal with CTC stain. Bacteria isolated from the UPW systems were mostly gram negative, and several groups seem to be indigenous for all of the UPW production systems studied. These included Ralstonia pickettii, Bradyrhizobium sp., Pseudomonas saccharophilia, and Stenotrophomonas strains. These bacteria constituted a significant part of the total number of isolated strains (≥20%). Two sets of primers specific to R. pickettii and Bradyrhizobium sp. were designed and successfully used for the detection of the corresponding bacteria in the concentrated UPW samples. Unexpectedly, nifH gene sequences were found in Bradyrhizobium sp. and some P. saccharophilia strains isolated from UPW. The widespread use of nitrogen gas in UPW plants may be associated with the presence of nitrogen-fixing genes in these bacteria.

Many industries suffer from the microbial contamination of ultrapure water (UPW). These include the semiconductor, pharmaceutical, food, and beverage industries. Within the semiconductor industry, ultrapure water is utilized in the final rinsing stage, and the presence of even a single bacterial cell and/or the products of cellular degradation, can severely compromise the quality of the final product (33, 39).

The industrial production of UPW is a complex multistep process, which involves two major stages referred to as pretreatment and polishing (Fig. (Fig.1).1). A variety of steps are included in many UPW production systems (e.g., filtration, UV light treatment, heat treatment, and ozonation) to remove and destroy bacteria. In particular, treatment with UV254 light and ozonation are present in some parts of a facility solely to prevent microbial contamination. Nitrogen gas is often used instead of air above stored UPW to prevent carbon dioxide and oxygen from dissolving in the water. It is imperative that UPW is kept carbon dioxide-free to prevent ionic loading on the mixed-bed ion-exchange resins, while the lowering of oxygen concentration should minimize bacterial growth (Fig. (Fig.1).1). Despite these precautions, piping, membranes, tanks, and other surfaces within the UPW system provide favorable places for bacterial adhesion and cell growth. The complete removal of contaminating microorganisms is considered to be nearly impossible (11, 20).

FIG. 1.
Schematic presentation of the typical UPW production system studied in this work. Direction of the water flow shown by arrows. Most of the components presented on the diagram are common to all five UPW systems investigated in this work, although the order ...

Although UPW contains less than part-per-billion quantities of inorganic and organic molecules, a group of microorganisms known as oligotrophs have adapted to these stringent conditions (22, 26). Many of these bacteria can excrete extracellular polysaccharides, allowing both adherence to surfaces and potential resistance to disinfection (14, 16, 37). The extracellular polysaccharide matrix acts as a diffusion barrier to nutrients and cellular products and allows nutrients from the flowing water to reach bacterial cells (7).

The biofilms present in UPW systems may be several cell layers thick (11, 20). The dead cells accumulating in biofilms may themselves be used as a carbon source by successive generations of bacteria. This phenomenon is often referred to as cryptic growth (29). The removal or disruption of biofilms in piping remains a challenge to UPW users.

While investigators have addressed the issue of the microbial contamination of UPW, few studies have been conducted to reveal the diversity of bacterial populations present in UPW. More importantly, except for the work by Pepper et al. (24, 25), most studies have been concerned only with bacterium assessment by agar plating techniques. As outlined previously (4, 18, 19), bacterial enumeration by such methods can lead to a vast underestimation of the actual levels of bacterial presence in various environments. It is generally accepted that gram-negative bacteria predominate in UPW (9, 17, 39), and it was shown that Pseudomonas species can be present in distilling and UPW systems (6, 13, 16).

Since high-purity water is widely used in many industries, this manufactured type of environment has acquired global importance. The investigation of bacterial diversity is essential for understanding of microbial populations inhabiting UPW. Such investigations will lead to characterization of the nutritional requirements of UPW bacteria and to the assessment of the surfaces used for biofilm formation. Identification of the main bacterial groups contaminating UPW may lead to the construction of probes for the detection and real-time monitoring of biocontamination.

We investigated here the diversity of bacterial communities in two university and four industrial UPW systems. More emphasis was placed on the microorganisms found in the polishing loops (especially distribution lines) of the systems. Oligonucleotide probes specific to the main bacterial species inhabiting UPW were designed. These probes were then successfully used to directly detect bacteria present in five different UPW plants. Pseudomonas, Ralstonia, and Bradyrhizobium species were shown to be present in most of the analyzed UPW systems.


Media and bacterial strains.

R2A media (28) was used in this work for growth and analysis of the bacterial strains present in UPW. This medium is recommended by American Society for Testing and Materials (ASTM) (1, 2) for testing UPW quality and is therefore widely used in industries. All bacterial strains investigated in this study were isolated from water samples obtained at different UPW plants. Three of the six plants analyzed in this study are used for production of UPW in semiconductor manufacturing processes. Designation of the UPW systems analyzed in this work is as follows: UPWS-1 (University of Arizona experimental UPW system, pilot scale), UPWS-2 (University of Arizona experimental UPW system number 2, bench scale), UPWS-3 (industrial UPW system number 1), UPWS-4 (industrial UPW system number 2), UPWS-5 (industrial UPW system number 3, not semiconductor industry), and UPWS-6 (industrial UPW system number 3, not semiconductor industry). Figure Figure11 shows a schematic of the common parts of the analyzed UPW systems. Nitrogen was used in polishing loops of UPWS-1, UPWS-3, UPWS-5, and UPWS-6.

UPW sample collection.

The procedure for UPW sample collection was described in detail previously (18). Briefly, before taking water samples, the ports' exteriors were cleaned with 70% ethanol, and water was allowed to flow for 3 min (50 ml/min). Samples were collected into sterile Whirlpak tubes and analyzed within 24 h or sooner depending on the location of the UPW plant.

For the epifluorescence microscopy analysis, 10 liters of water was filtered through a black polycarbonate membrane (Nuclepore [Corning], 0.2-μm pore size) as described by McAlister et al. (18).

For detection of 16S rRNA gene sequences in UPW by PCR (direct PCR) bacterial cells from the UPW were concentrated onto polycarbonate membranes (0.1-μm pore size) as described above. The membrane was aseptically removed from the filter holder and transferred to a sterile polypropylene centrifuge tube (50 ml) containing 10 ml of double-filter-sterilized UPW. This was incubated at 25°C (180 rpm) overnight. After incubation, the cells were concentrated by centrifugation (8,000 rpm, 15 min) and resuspended in 1 ml of double-filter-sterilized UPW. Concentrated water samples were used directly in the PCRs.

Epifluorescence microscopy.

Cyanotolyl tetrazolium chloride (CTC) and DAPI (4′,6′-diamidino-2-phenylindole) staining techniques were based on the procedure of Pyle et al. (27). All staining solutions were prepared in UPW and double filter sterilized prior to use. Both CTC and DAPI were obtained from Polysciences (Warrington, Pa.). Stained membranes were examined with an epifluorescence microscope (Olympus BH-2) by using the filter combinations described previously (12). A minimum of 20 microscope fields (using an ocular grid of known dimensions) were counted for each membrane. The DAPI count reflected the total number of bacterial cells present, while the CTC count represented the number of cells with the potential for respiration.

Determination of 16S rRNA gene sequence.

Total DNA was isolated from bacterial cells grown to an optical density at 600 nm of between 0.8 and 1.0. After centrifugation the cells were resuspended in 90 μl of 50 mM Tris-HCl buffer (pH 8.0) containing sucrose (6.2% [wt/vol]) and EDTA (12 mM). Immediately, 20 μl of lysozyme (30 mg/ml; Sigma) was added. After 15 min of incubation at 37°C, 15 μl of sodium dodecyl sulfate (SDS; 20% [wt/vol]) was added, and the mixture was incubated at 37°C for 1 h. The preparation was then extracted with an equal volume of phenol and then with phenol-chloroform (1/1 [vol/vol]). DNA was then precipitated with ethanol, washed once, and finally resuspended in 60 μl of Tris-EDTA buffer (30).

An almost-complete 16S rRNA gene was amplified by PCR with the following universal primers, described by Pascual et al. (21): forward (5′-AGAGTTTGATCCTGGCTCAG, positions 8 to 27 [Escherichia coli numbering]) and reverse (5′-AAGGAGGTGATCCAGCCGCA, positions 1541 to 1522). Amplification of the 16S rRNA genes was done with Taq+ DNA polymerase (Stratagene) in a buffer supplied by the manufacturer. Reactions were carried out in volumes of 25 μl with deoxynucleoside triphosphates at 200 μM concentrations, primers at 0.15 μM each, DNA at 100 to 200 ng, and Taq+ at 0.5 U per reaction. The following temperature profile was used: denaturation at 95°C for 3 min, followed by 30 cycles of 94°C for 40 s, 60°C for 30 s, and 72°C for 1 min. The amplification reactions were carried out by using a Perkin-Elmer DNA Thermal Cycler 480. The PCR products were purified by using GFX PCR DNA and Gel Band Purification Kit (Pharmacia Biotech). When the cloning of PCR fragments was required, Pfu polymerase was used for blunt-end generation, and the resulting products were cloned into the SmaI site of the pUC19 vector plasmid. Plasmid DNA was isolated by standard procedures (30).

Purified PCR products or plasmid DNA were used in sequencing reactions with the Taq Dye-Deoxy Terminator Cycle Sequencing Kits (Applied Biosystems and Beckman). The primers used for PCR amplification were also employed for sequencing. Additional sequencing primers were designed on the basis of conserved regions of eubacterial 16S rRNA genes (35), as well as on the basis of preliminary information obtained by sequencing of UPW isolates. The forward primers (E. coli numbering) used in this work were as follows: LK256, 5′-GGTTAAGTCCCGCAACGA-3′ (positions 1364 to 1381); LK258, 5′-CTCCTACGGGAGGCAGCA-3′ (positions 339 to 356); LK272, 5′-TGCCAGCAGCCGCGGTA-3′ (positions 516 to 532); and LK274, 5′-AGCAAACAGGATTAGATACC-3′ (positions 1053 to 1072). The reverse primers were as follows: LK257, 5′-TCGTTGCGGGACTTAACC-3′ (positions 1381 to 1364); LK266, 5′-ACTGCTGCCTCCCGTAGGA-3′ (positions 358 to 340); LK273, 5′-TACCGCGGCTGCTGGCA-3′ (positions 532 to 516); and LK275, 5′-GGCGTGGACTACCAGGGTA-3′ (positions 1087 to 1069). The nucleotide sequences of both strands were determined by using automatic sequencers (Applied Biosystems model 373A and the Beckman CEQ 2000 DNA Analysis System).

Editing and initial analysis of the sequences was performed by using the DNASIS (Hitachi) software package. Searches for nucleotide and amino acid sequence similarities were done by using the FASTA and BLAST programs (23) and the EMBL and GenBank databases.

Alignments of the sequences were performed by using the CLUSTALW program (32). Phylogenetic analysis of the alignment was done by using the PHYLIP (version 3.57c) package (10) and the TREECON program (34). For the PHYLIP analysis, bootstraps were obtained with the SEQBOOT program (100 data sets were generated). Parsimony analyses were done with the DNAPARS programs with ordinary parsimony and randomized input order of the sequences. For the analyses with the TREECON program, Tajima and Nei correction (31) was used, and trees were generated by neighbor joining.

PCR detection of bacterial contamination in UPW.

Concentrated UPW samples (3 μl) were used directly in PCRs essentially as was described by Pepper et al. (25). Conditions for PCR were as described above, except the cycling profile used was as follows: denaturation at 95°C for 5 min, followed by 35 cycles of 94°C for 40 s, 55°C for 30 s, and 72°C for 1 min. The preparations were then analyzed by 1.2% agarose gel electrophoresis. For the detection of bacterial contamination in the water samples, three sets of primers were used: universal eubacterial primers (see above), primers designed for Bradyrhizobium sp. (forward primer LK288 [5′-CGTAAAGGGTGCGTAGGCGGGTCTTTA-3′], positions 509 to 535; reverse primer LK289 [5′-CCCTTTCGGTTAGCGCACCGTCTT-3′, positions 1388 to 1365; the estimated fragment size is 880 bp), and primers designed for Ralstonia pickettii (forward primer LK290 [5′-TGTCCGGAAAGAAATGGCTCTGG-3′], positions 416 to 438; reverse primer LK291 [5′-CTAACTACTTCTGGTAAAGCCCAC-3′], positions 1413 to 1390; the estimated fragment size is 975 bp). These sets of primers were designed as specific to bacterial strains found in UPW only and therefore should not be considered species specific (e.g., LK288, apart from Bradyrhizobium sp., is homologous to Nitrobacter spp. and some other bacteria; LK289 is specific only to Bradyrhizobium spp. but not to all Bradyrhizobium strains; the same limitations apply to R. pickettii primers).

Detection of nifH genes.

To detect the genes responsible for nitrogen fixation in UPW bacteria, two previously described primers (38) were used: primer 19F (5′-GCIWTYTAYGGIAARGGIGG-3′) and primer 407R (5′-AAICCRCCRCAIACIACRTC-3′).

Prevention of contamination.

Because of the very low numbers of bacterial cells present in UPW, possible contamination of samples represents an important issue. For PCR analysis of UPW, the precautions described by Pepper et al. (25) were followed. When bacteria were isolated by plating them on R2A medium, the controls included swabs from the port exterior, autoclaved UPW samples, plating the bacteria present in the surrounding environment, and plating the bacteria from the water entering the UPW system.


Analysis of the University of Arizona UPW system (UPWS-1) was central to this investigation. Bacterial contamination of this system was regularly monitored for 2.5 years. The results of these surveys have been in part reported elsewhere (18), but no species identification was reported. Subsequently, in a comparative study we analyzed the bacterial communities present in UPWS-1, UPWS-2, UPWS-3, UPWS-4, UPWS-5, and UPWS-6.

Isolation and characterization of UPW bacteria.

Previous analysis of UPWS-1 showed that the majority of UPW strains are facultative oligotrophs (with no obligatory oligotrophs found). They grew equally well on the full-strength R2A medium and its dilutions (18). R2A media therefore were used in this study. The influence of incubation times on the enumeration of CFU present in UPWS-1 has been reported previously (18). In accordance with those results, plate counts on R2A medium and bacterial isolations were conducted after 4 weeks of incubation at 25°C. All of the isolated strains were purified by using the same media. Preliminary characterization showed that the majority of the strains isolated were gram-negative bacteria. All isolations and bacterial counts were conducted in aerobic conditions as recommended by ASTM (1, 2). Our preliminary experiments also failed to produce bacterial growth under anaerobic conditions (results not shown).

After initial characterization of the isolated strains, corresponding 16S rRNA gene sequences were obtained. Sequences of at least 900 bp were determined, and for every group of closely related sequences (homology of >99%), an almost complete 16S rRNA gene sequence (1,400 to 1,500 bp) was obtained for at least one bacterial isolate. The results of phylogenetic analysis of bacteria present in UPW obtained from five different plants are shown in Table Table1,1, and a phylogenetic tree of the main bacterial strains isolated from UPW (UPWS-1) is presented in Fig. Fig.2.2. All of the UPW samples analyzed were collected from ports situated in the final (polishing loop) parts of the corresponding UPW systems (mostly the distribution line, indicated in Fig. Fig.11).

FIG. 2.
Phylogenetic analysis of the 16S rRNA genes from strains isolated from University of Arizona UPW System (UPWS-1). The tree was obtained by the neighbor-joining approach by using the TREECON program. Similar phylogenies were obtained when parsimony analysis ...
Bacterial strain identification on the basis of 16S rRNA gene analysis in tested UPW systemsa

The results indicate that some bacterial species are present in most or all of the UPW systems analyzed. These bacteria were strains most closely related to R. pickettii (found in four of six analyzed UPW systems), strains related to several Bradyrhizobium sp. (present in all but one of the analyzed UPW systems), and Pseudomonas saccharophilia (found in three UPW systems). These bacterial species constituted a significant part of the total number of isolated strains (≥20%) (Table (Table1).1). It is important to note that Bradyrhizobium strains isolated from UPW required at least 7 days to develop visible colonies on R2A medium (25 to 30°C). Under the same incubation conditions, most Ralstonia and Pseudomonas strains grew within 1 to 2 days. Bradyrhizobium strains have not previously been reported in these types of systems.

A phylogeny of the typical representatives of these bacterial species is given in Fig. Fig.2.2. Strains related to R. pickettii and Bradyrhizobium sp. were first detected in UPWS-1 and continued to appear in every survey conducted on this UPW system. P. saccharophilia strains were isolated from UPWS-1 in February 1999, prior to the plant being sanitized, and have not been found in UPWS-1 since then. This bacterium was also isolated in significant numbers at UPWS-4 and UPWS-6. None of the above bacterial types were detected in the control samples (i.e., in incoming water or air samples taken on the sites). The other types of bacteria detected mainly in UPWS-1 samples were Stenotrophomonas, Ralstonia, and Flavobacterium spp. This is most likely because this system was analyzed much more completely and repeatedly. Some other bacterial strains were characteristically present in lower numbers or did not appear to be present in more than one UPW system and, in some cases, were also detected in incoming water (e.g., Mycobacterium and Bacillus spp.).

Assessment of the extent of bacterial contamination of UPW.

In addition to identification of bacterial types described above, bacterial numbers present in the UPW system were determined. We report here the analysis of bacterial contamination in the polishing parts of UPW systems and compare different UPW systems. UPW samples taken from the ports located at the final stages of water treatment were analyzed. Bacteria present in UPW were enumerated by both agar plate counts and direct counts by using epifluorescence microscopy. The results of the analysis are presented in Table Table2.2. It is important to note that UPWS-1, UPWS-2, UPWS-3, UPWS-4, and UPWS-5 systems showed approximately the same numbers of bacteria when assessed by plate counts after incubation for 4 weeks (somewhat higher for UPWS-6). Assessment of bacterial presence in UPW by epifluorescence microscopy (CTC and DAPI staining) showed significantly higher numbers (10 to 100 times more bacterial cells were detected) for UPWS-1, UPWS-2, UPWS-3, UPWS-4, and UPWS-5 systems (Table (Table2).2). Only the bacterial numbers obtained for UPWS-5 (by CTC staining) corresponded to those by plate counts. Somewhat higher level of UPWS-6 contamination (as shown by plate counts) may point to the presence of a higher percentage of organics in the water. A significant proportion of the bacteria present in UPW (50 to 90%) appeared to be composed of nonviable cells. Although the lack of CTC signal does not necessary means that an organism is nonviable, the ratio of DAPI to CTC counts may serve as a preliminary assessment of percentage of nonviable bacteria in the populations. In UPWS-4, the number of nonviable cells is particularly high (Table (Table22).

Assessment of bacterial contamination of UPWa

It is noteworthy that bacterial numbers did not vary significantly in samples obtained from various points of the polishing section of the UPW production systems tested (although a fivefold decrease in bacterial numbers may be noted in UPWS-3 after thermal treatment of UPW) (Table (Table22).

It was previously shown that bacteria in UPW systems grow as biofilms on the inner surfaces of pipings (11, 20). To confirm the origin of planktonic bacteria investigated in this work, swabs were taken in various parts of the polishing loop (UPWS-1), and bacteria thus collected were identified as described above. The analysis of the samples isolated from swabs confirmed the presence of bacterial biofilms on the inner surfaces of UPW system. No new genera or species were detected by this analysis (i.e., bacteria isolated showed the same identities as those isolated from UPW samples; Table Table1).1). This analysis indicates that planktonic bacteria species isolated from UPW samples represent true diversity of bacterial populations in UPW systems.

Detection of UPW bacteria by PCR analysis.

Detection of contaminating bacteria by direct PCR of UPW samples was first reported by Pepper et al. (24, 25). In the present study, we used PCR for the detection of bacterial contamination in different UPW plants employed by the semiconductor and other manufacturers. Two sets of primers specific to R. pickettii and Bradyrhizobium sp., as well as primers universal for eubacteria, were used. The specificity of the primers was tested in control experiments involving various laboratory bacterial strains and the strains isolated from UPW; the identities of PCR products obtained with specific primers were also confirmed by sequencing. The results of direct PCR analysis of UPW are presented in Table Table3.3. These results confirmed the possibility of detection of bacterial contamination of UPW by direct PCR analysis. The results obtained with specific primers correspond to those obtained by identification of the isolated bacteria (Table (Table1).1). In some cases there were no PCR products obtained (UPWS-4, “After UV254” [see Table Table3]),3]), which is probably due to the very low numbers of bacterial cells present in particular UPW samples. Primers specific to R. pickettii and Bradyrhizobium sp. allowed detection of the corresponding bacteria in UPW. It is worth noting that, apart from R. pickettii, the primers designed may target several other species. Although these primers always behaved as specific in our experiments with UPW samples, such a possibility should be taken into account when different UPW systems are analyzed.

PCR detection of bacteria in UPWa

PCR products obtained with the universal bacterial primers from the UPWS-1, UPWS-3, UPWS-4, and UPWS-5 samples (Table (Table3)3) were cloned in the pUC18 vector, and partial sequences of the two or three insertions were obtained in each case (ca. 500 bp). The bacterial species identified corresponded to those presented in Table Table1,1, i.e., sequences obtained from the UPWS-1, UPWS-3, and UPWS-5 samples were identified as belonging to R. pickettii and sequences from UPWS-4 were identified as belonging to Sphingomonas sp. (results not shown).

Detection of nifH genes in bacteria isolated from UPW.

A number of Bradyrhizobium strains were isolated from various analyzed UPW systems. Since nitrogen is used in most of these systems to reduce O2 and CO2, its presence may be instrumental in supporting bacterial growth within UPW systems. Although UPW systems have very low overall level of carbon and organic compounds, locally (cryptic growth in biofilms) that level may be sufficient to supply energy for nitrogen fixation. As a first approach to investigate this hypothesis, the distribution of nifH gene sequences in the UPW bacterial community has been analyzed. All bacterial strains isolated from UPWS-1 and a number of isolates obtained from the four other systems were analyzed for the presence of nifH genes. It was shown that nifH gene sequences are present in ca. 60% of Bradyrhizobium strains isolated from UPW. nifH genes were also detected in all four analyzed P. saccharophilia strains. Although the presence of nifH genes in Bradyrhizobium is well documented (38), they are more rarely found among Pseudomonas species (3, 5). No other bacterial isolates analyzed showed the presence of nifH genes. It should be noted that the presence of nifH genes does not necessarily mean the activity of nitrogenase, since this enzyme is regulated at both pre- and posttranslational levels (8).


The bacterial diversity within the UPW systems primarily employed in the semiconductor industry has been investigated. Six UPW systems were analyzed, two smaller university systems and four full-size industrial plants, that were located in geographically diverse areas of the world. Because the UPW systems varied in size and location, the results obtained are considered to be characteristic for UPW production in general. The samples analyzed here were obtained from the ports located in the polishing sections of UPW systems; hence, the bacterial communities investigated may have a significant impact on the quality of UPW used in the final (rinsing) stages of semiconductor production.

Five UPW systems showed approximately the same level of bacterial contamination as assessed by different methods. Considering the different locations and sizes of the analyzed UPW systems, the bacteria detected may be considered as indigenous populations typically found in these systems. A comparison of the UPW bacterial contamination by plate counts to that of DAPI and CTC staining detected a significant underestimation of bacterial presence by the former (with the exception of UPWS-6; similar bacterial numbers were detected by plate counts and epifluorescence microscopy). Similar results have already been reported for drinking water and UPW analysis (4, 18, 19). It is important to note that detection and estimation of bacteria within the semiconductor industry still relies heavily on direct cultivation and plating techniques according to ASTM standards (1, 2). Thus, the results of such procedures may significantly underestimate the extent of the problem.

The bacteria isolated from the UPW systems were mostly gram negative, and several groups seem to be indigenous for UPW production systems. These included R. pickettii, Bradyrhizobium, P. saccharophilia, Stenotrophomonas, and Ralstonia strains. It is worth noting that identification solely on the basis of 16S rRNA gene sequences (this work) is not sufficient for drawing a reliable distinction between species. It is essential to note that UPWS-1 was analyzed far more rigorously than the other four systems; UPW samples were taken from UPWS-1 on a bimonthly basis, and bacteria present were analyzed by plating, epifluorescence microscopy, and PCR. UPW samples from other plants were obtained once or twice during the same period.

From the analysis of UPWS-1, it became clear that the above groups of bacteria represent the most important parts of the bacterial community in polishing and distribution parts of the system, whereas various other bacterial species were found in the sections of the plant located upstream of the final UV lamps and filters (Table (Table1).1). Analysis of the other UPW plants showed that the bacterial groups isolated from UPWS-1 were also the main bacteria inhabiting other UPW environments. R. pickettii strains were found in four of the six analyzed UPW systems, and strains identified as mostly close to Bradyrhizobium sp. were present in all but one UPW systems. Previous research showed that various representatives of the genus Pseudomonas are present in UPW (6, 16, 22, 33). P. aeruginosa and Burkholderia cepacia were shown to contaminate a water-distilling system (13). R. pickettii strains were also reported present among many others species in UPW (6). However, there were no Bradyrhizobium strains detected in UPW previously, and no attempts have been made to identify the typical bacterial strains inhabiting various UPW production systems. Failure to isolate Bradyrhizobium strains from UPW in previous reports may be due to the relatively slower growth of these bacteria on standard R2A media; ASTM guidelines (1, 2) recommend growing the bacteria (for isolation from UPW) for 48 to 72 h. Our experiments showed that this period of incubation is insufficient for growth of Bradyrhizobium sp. present in UPW systems (18).

It is important to emphasize that very little variation in bacterial counts was observed when different parts of UPW polishing loops were analyzed. This observation questions the effectiveness of some stages of antibacterial treatment employed in modern UPW production: in particular, UV254 treatment seems to have little effect on the bacterial numbers present in water. These findings may be better understood in conjunction with the nature of bacterial growth in UPW systems, which according to most available evidence occurs in the form of biofilms attached to inner surfaces (11, 20). If we take into consideration the results of the present study, it may be suggested that each part of the UPW production system (separated from others by filters, UV-units, etc.) possesses its own bacterial population (biofilm) relatively independent from others present in the same system. Correspondingly, planktonic bacteria detected in UPW represent cells detached from biofilms. Although this suggestion seems reasonable and corresponds with the results obtained, further experimental work may be needed for its confirmation. It is worth noting that Bradyrhizobium strains were detected in all plants where nitrogen had been used; however, the same group of bacteria was also isolated from the UPWS-2, which does not contain nitrogen. The industrial plant seemingly free from Bradyrhizobium sp. did not use nitrogen in its system (UPWS-4).

Discovery of nifH genes in Bradyrhizobium strains is not surprising by itself, but when taken in conjunction with the spread of this bacterial group in UPW systems, it may suggest that nitrogen contributes to cell maintenance and growth in these systems. The presence of nifH sequences in P. saccharophilia is somewhat more unusual. However, a few examples of nitrogen-fixing Pseudomonas strains have been reported (5, 15, 36), and a nitrogen-fixing strain of P. saccharophilia ATCC 15946 has also been reported (3). It is important to stress that finding bacterial strains with nif genes deserves further investigation, since it provides the first evidence that the widespread use of nitrogen in UPW production systems may contribute to bacterial contamination.

Identification of the typical bacterial strains inhabiting UPW systems allowed us to design oligonucleotide primers specific to two main bacterial groups. PCR experiments conducted with UPW samples demonstrated the possibility for detection of R. pickettii and Bradyrhizobium sp. in UPW. A 100- to 1,000-fold concentration of water samples was needed for such detection, since bacterial numbers in typical UPW taken from the system were too low to allow direct PCR detection of bacterial 16S rDNA sequences. The use of specific (as well as universal) primers for the detection of the bacterial contamination in UPW systems may be considered useful for the preliminary assessment of bacterial presence in UPW.

In conclusion, it should be said that certain bacterial populations appear common to many industrial UPW systems and are represented mostly by gram-negative strains. Several bacterial species were found (Pseudomonas, Ralstonia, and Bradyrhizobium spp.) that seem to be indigenous to an oligotrophic UPW environment.


We acknowledge the support of the NSF Industry/University Cooperative Research Center Program and of the Industrial Research and Technology Unit Northern Ireland START Programme Grant under the international “TIE” Project, “Microbiocontamination in Ultrapure Water,” involving researchers at the University of Arizona, The Queen's University of Belfast, SUNY at Buffalo, and NJIT.

We also thank four industrial sites that allowed us to sample their UPW systems and Jon Sjogren for help with the collection of UPW samples.


1. American Society for Testing and Materials. 2000. Standard test methods for microbiological monitoring of water used for processing electron and microelectronic devices by direct pressure tap sampling valve and by the presterilized plastic bag method (F-1094), p. 287-290. In Annual Book of ASTM Standards. American Society for Testing and Materials, West Conshohocken, Pa.
2. American Society for Testing and Materials. 2000. Standard guide for ultrapure water used in the electronics and semiconductor industry (D-5127), p. 495-499. In Annual Book of ASTM Standards. American Society for Testing and Materials, West Conshohocken, Pa.
3. Barraquio, W. L., B. C. Parde, Jr., I. Watanabe, and R. Knowles. 1986. Nitrogen fixation by Pseudomonas saccharophila Doudoroff ATCC 15946. J. Gen. Microbiol. 132:237-241.
4. Byrd, J. J., H. S. Xu, and R. R. Colwell. 1991. Viable but not culturable bacteria in drinking water. Appl. Environ. Microbiol. 57:875-878. [PMC free article] [PubMed]
5. Chan, Y.-K., W. L. Barraquio, and R. Knowles. 1994. N2-fixing pseudomonads and related soil bacteria. FEMS Microbiol. Rev. 13:95-118.
6. Clancy, J. L., and L. Cimini. 1991. Improved method for recovering bacteria from HPW water. Ultrapure Water 8:25-37.
7. Cooksey, K. E. 1992. Extracellular polymers in biofilms, p. 137-147. In L. F. Melo, T. R. Bott, M. Fletcher, and B. Capdeville (ed.), Biofilms: science and technology. Kluwer Academic Publishers, Dordrecht, The Netherlands.
8. Dean, D. R., and M. R. Jacobson. 1992. Biochemical genetics of nitrogenase, p. 763-834. In G. Stacy, R. H. Burris, and H. J. Evan (ed.), Biological nitrogen fixation. Chapman and Hall, New York, N.Y.
9. Favero, M. S., N. J. Petersen, L. A. Carson, W. W. Bond, and S. H. Hindman. 1975. Gram-negative water bacteria in hemodialysis systems. Health Lab. Sci. 12:321-334. [PubMed]
10. Felsenstein, J. 1997. PHYLIP (Phylogeny Inference Package), version 3.57c. Department of Genetics, University of Washington, Seattle.
11. Henley, M. 1992. Sanitization or sterilization? It depends on the final use for the high-purity water. Ultrapure Water 9:15-21.
12. Kawai, M., N. Yamaguchi, and M. Nasu. 1999. Rapid enumeration of physiologically active bacteria in purified water used in the pharmaceutical manufacturing process. J. Appl. Microbiol. 86:496-504. [PubMed]
13. Kayser, W. V., K. C Hickman, W. W. Bond, M. S. Favero, and L. A. Carson. 1975. Bacteriological evaluation of an ultra-pure water-distilling system. Appl. Microbiol. 30:704-706. [PMC free article] [PubMed]
14. LeChevallier, M. W., T. S. Hassenauer, A. K. Camper, and G. A. McFeters. 1984. Disinfection of bacteria attached to granular activated carbon. Appl. Environ. Microbiol. 48:918-923. [PMC free article] [PubMed]
15. Line, M. A. 1997. Nitrogen-fixing consortia associated with the bacterial decay of a wooden pipeline. Lett. Appl. Microbiol. 25:220-224.
16. Martyak, J. E., J. C. Carmody, and G. R. Husted. 1993. Characterizing biofilm growth in deionized ultrapure water piping systems. Microcontamination 1993:39-44.
17. Matsuda, N., W. Agui, T. Tougou, H. Sakai, K. Ogino, and M. Abe. 1996. Gram-negative bacteria viable in ultrapure water isolated from ultrapure water and effect of temperature on their behavior. Colloids Surf. B Biointerfaces 5:279-289.
18. McAlister, M. B., L. A. Kulakov, M. J. Larkin, and K. L. Ogden. 2001. Analysis of bacterial contamination in different sections of a high-purity water system. Ultrapure Water 18:18-26.
19. McCoy, W. F., and B. H. Olson. 1986. Relationship among turbidity, particle counts, and bacteriological quality within water distribution lines. Water Res. 20:1023-1029.
20. McFeters, G. A., S. C. Broadway, B. H. Pyle, K. K. Siu, and Y. Egozy. 1993. Bacterial ecology of operating laboratory water purification systems. Ultrapure Water 10:32-37.
21. Pascual, C., P. A. Lawson, J. A. E. Farrow, M. N. Gimenez, and M. D. Collins. 1995. Phylogenetic analysis of the genus Corynebacterium based on 16S rRNA gene sequences. Int. J. Syst. Bacteriol. 45:724-728. [PubMed]
22. Patterson, M. K., G. R. Husted, A. Rutkowski, and D. C. Mayette. 1991. Isolation, identification and microscopic properties of biofilms in high-purity water distribution systems. Ultrapure Water 8:18-23.
23. Pearson, W. R., and D. J. Lipman. 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85:2444-2448. [PMC free article] [PubMed]
24. Pepper, I. L., K. L. Josephson, R. L. Bailey, M. D. Burr, and S. D. Pillai. 1993. A rapid and systematic analytical method for measuring bacterial contaminants in ultrapure water, p. 50-62. In Proceedings of the Semiconductor Pure Water and Chemicals Conference, 2 to 4 March 1993, Santa Clara, Calif.
25. Pepper, I. L., K. L. Josephson, R. L. Bailey, M. D. Burr, S. D. Pillai, D. L. Tolliver, and S. Pulido. 1994. Measuring bacterial contamination in ultra pure water: a rapid analytical method. Microcontamination 1994:52-57.
26. Poindexter, J. S. 1981. Oligotrophy: fast and famine existence. Adv. Microb. Ecol. 5:63-89.
27. Pyle, B. H., S. C. Broadaway, and G. A. McFeters. 1995. Factors affecting the determination of respiratory activity on the basis of cyanoditolyl tetrazolium chloride reduction with membrane filtration. Appl. Environ. Microbiol. 61:4304-4309. [PMC free article] [PubMed]
28. Reasoner, D. J., and E. E. Geldreich. 1985. A new medium for the enumeration and subculture of bacteria from potable water. Appl. Environ. Microbiol. 49:1-7. [PMC free article] [PubMed]
29. Roszak, D. B., and R. R. Colwell. 1987. Survival strategies of bacteria in the natural environment. Microbiol. Rev. 51:365-379. [PMC free article] [PubMed]
30. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
31. Tajima, F., and M. Nei. 1984. Estimation of evolutionary distance between nucleotide sequences. Mol. Biol. E 1:269-285. [PubMed]
32. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680. [PMC free article] [PubMed]
33. Tsuchizaki, M. 1983. Measurement of microbes and particulates in ultrapure water and chemicals for the electronic industry. Semiconductor World 1983:1-20. (English Translation.)
34. Van de Peer, Y., and R. De Wachter. 1994. TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Comput. Appl. Biosci. 10:569-570. [PubMed]
35. Van de Peer, Y., S. Chapelle, and R. DeWachter. 1996. A quantitative map of nucleotide substitution rates in bacterial rRNA. Nucleic Acids Res. 24:3381-3391. [PMC free article] [PubMed]
36. Vermeiren, H., A. Willems, G. Schoofs, R. de Mot, V. Keijers, and W. Hai. 1999. The rice inoculant strain Alcaligenes faecalis A15 is a nitrogen-fixing Pseudomonas stutzeri. Syst. Appl. Microbiol. 22:215-224. [PubMed]
37. Vess, R. W., R. L. Anderson, J. H. Carr, W. W. Bond, and M. S. Favero. 1993. The colonization of solid PVC surfaces and acquisition of resistance to germicides by water microorganisms. J. Appl. Bacteriol. 74:215-221. [PubMed]
38. Ueda, T., Y. Suga, N. Yahiro, and T. Matsuguchi. 1995. Remarkable N2-fixing bacterial diversity detected in rice roots by molecular evolutionary analysis of nifH gene sequences. J. Bacteriol. 177:1414-1417. [PMC free article] [PubMed]
39. White, D. C., and M. W. Mittleman. 1990. Biological fouling of high purity waters: mechanisms and consequences of bacterial growth and replication, p. 150-171. In Proceedings of the Ninth Annual Semiconductor Pure Water Conference, 17 and 18 January 1990, Santa Clara, Calif.

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...