Ice is defined as a food and is frequently used in direct contact with food and beverages. Packaged ice is commercially produced and can be easily found in grocery and convenience stores. However, the quality and safety of packaged ice products is not consistent. The Packaged Ice Quality Control Standards manual (PIQCS) published by the International Packaged Ice Association provides the quality and processing standards for packaged ice produced by its members. Packaged ice produced on the premise of stores (on-site packaged ice) is not required to be in compliance with these standards. In this study, packaged ice produced by manufacturing plants or by in-store bagger (ISB) machines and on-site packaged ice were compared for their microbiological quality and microbial diversity. Our results revealed that 19% of the 120 on-site packaged ice samples did not meet the PIQCS microbial limit of 500 CFU/mL (or g) and also the absence of coliforms and Escherichia coli. Staphylococci were found in 34% of the on-site packaged ice samples, most likely through contamination from the packaging workers. None of the ISB and manufactured packaged ice samples had unacceptable microbial levels, and all were devoid of staphylococci. Salmonella was absent in all samples analyzed in this study. Microbial community analysis of ice based on 16S/18S rRNA targeted sequencing revealed a much higher microbial diversity and abundance in the on-site packaged ice than in the ISB ice. Proteobacteria, especially Alphaproteobacteria and Betaproteobacteria, were the dominant bacterial groups in all samples tested. Most of these bacteria were oligotrophic; however, a few opportunistic or potential pathogens were found at low levels in the on-site packaged ice but not in the ISB packaged ice. The types of microbes identified may provide information needed to investigate potential sources of contamination. Our data also suggest a need for enforcement of processing standards during the on-site packaging of ice.

Ice is a staple in American households and is produced on an industrial scale for many applications, including beverage mixing, food processing and storage, and chemical manufacturing. For consumption, ice is produced in shapes such as nuggets and tubes, typically packaged in 2- to 5-kg bags, and sold in grocery and convenience stores. Packaged ice can be produced by various methods and in various settings, including commercial-scale manufacturing plants (manufactured ice), self-sufficient in-store bagger machines (ISBs), and on convenience store premises (on-site packaged ice). According to the International Packaged Ice Association (IPIA) (26), approximately 2 billion bags of packaged ice are sold from retail, wholesale, and vending producers each year in the United States, and among them, 800 million bags (40%) are packaged on store premises and 200 million bags (10%) are from vending machine sales. Packaged ice is considered food and therefore is under the jurisdiction of the U.S. Food and Drug Administration (FDA). However, safe ice manufacturing and packaging practices are often overlooked, and ice continues to contribute to foodborne outbreaks worldwide (9). The IPIA publishes “The PIQCS Manual (Packaged Ice Quality Control Standards)” (25) as mandatory practices for its members. The PIQCS specifies the good manufacturing practices and quality standards for packaged ice. For example, the PIQCS microbial limit for the total plate count (TPC) is 500 CFU/mL (or 500 CFU/g), and the total coliform level (including Escherichia coli) should be zero in 100 mL. According to the IPIA, approximately 800 million bags of the manufactured ice sold in the United States each year are produced by commercial ice-making companies that comply with PIQCS packaged ice processing standards; however, the remaining 1.2 billion bags of packaged ice each year are not produced under such standards, especially bags of ice sold directly from vending machines or packaged on site in retail stores with an ice machine (38).

Few studies on the quality of ice have been reported. In a study of 22 packaged ice samples collected in 1991 in Iowa (39), 8 samples (36%) exceeded the TPC limit established by the IPIA, 15 (68%) were positive for coliforms, and 13 (59%) were positive for mold species. All of the unsatisfactory samples did not comply with the IPIA's PIQCS. The bacteria isolated in that study were frequently opportunistic pathogens, such as Pseudomonas, Acinetobacter, Staphylococcus, Alcaligenes, and Corynebacterium. Chemical standards were satisfactory, except for five samples that were outside of the standard pH range of 6.5 to 8.5 (39). The analysis of ice quality in Florida reported in 1999 revealed coliforms in 13.5% of on-site packaged ice and 3.6% of manufactured ice, but the chemical and physical qualities were comparable between the two types of packaged ice products (47). A recent study in Georgia revealed that manufactured ice was much higher in microbiological and chemical quality than was on-site packaged ice from retail stores and vending machines (38). These researchers reported that 6.4 and 37% of the 250 on-site packaged ice samples had unsatisfactory TPCs and coliform levels, respectively. The pathogen Salmonella was found in one ice sample from a food service establishment, and Enterobacter agglomerans, a potential pathogen, was present in ice from a self-service vending machine (38). In all three of these studies, on-site packaged ice quality was less than satisfactory, mainly because of microbiological contamination. Additional data from different geographical regions are needed to validate these observations, especially in areas of high population density such as California; approximately 2.3 billion lb (1 billion kg) of manufactured ice, totaling over 340 million bags, is produced each year in California (27). Studies should also include samples of packaged ice made by ISBs; this type of sample has not been investigated previously.

In the present study, microbiological quality indicators (including TPCs and levels of coliforms and E. coli) were measured and compared among the manufactured, ISB, and on-site packaged ice samples collected from various areas in southern California. Among the pathogens, Salmonella was chosen as a target because of its hardiness in the food system and its association with major foodborne outbreaks (6, 12), and Staphylococcus was used as an indicator for contamination associated with food handlers (17, 34). Next generation sequencing was used to explore microbial diversity in selected samples. To our knowledge, this is the first report of a comprehensive microbial community analysis of ice products. Our results revealed that on-site packaged ice was more variable than other types of ice in terms of microbiological quality and contained higher levels and more diverse microbial communities.

Sample collection

A total of 156 packaged ice samples were collected from stores in southern California. The samples were grouped into manufactured, ISB, and on-site packaged ice, based on the production process. The 120 on-site packaged ice samples were further grouped into four geographical areas (Table 1). The on-site packaged ice samples were collected from convenience stores from six counties in southern California: Los Angeles, Orange, San Bernardino, Riverside, San Diego, and Imperial. These counties were further grouped into four areas based on geographical location and population density (Table 1). Thirty samples were collected from each geographical area, for a total of 120 on-site packaged ice samples. On-site packaging was verified by the presence of an ice making machine within the store and confirmation by the store employees. The size of the bags was 2.2 to 4.5 kg each. Six bags from each of the two brands of manufacturer-packaged IPIA-compliant ice samples were collected from retail stores. Six ice samples were also collected from different locations for each of the four ISB brands, for a total of 24 ISB samples. Two of the four ISB brands were IPIA compliant.

TABLE 1.

Microbiological quality of 156 packaged ice samples collected in southern California

Microbiological quality of 156 packaged ice samples collected in southern California
Microbiological quality of 156 packaged ice samples collected in southern California

Sample preparation

Immediately after collection, each ice bag was placed in a large, sterile 5-kg Whirl-Pak bag (Nasco, Modesto, CA) to avoid leakage and cross-contamination. The samples were labeled and transported to the laboratory in a large cooler. Upon arrival at the laboratory, ice samples were aseptically poured into new 5-kg Whirl-Pak bags and allowed to thaw overnight at room temperature. After complete melting of the ice, samples were thoroughly mixed before being used for individual tests. Sample preparation was completed within 24 h of collection.

Microbiological analyses

TPCs, total coliform and E. coli (CEC) counts, and staphylococci counts were performed. TPCs were performed using the SimPlate TPC (BioMedix, Pomona, CA) (51). CEC counts were determined using membrane filtration and MI medium (BD, Franklin Lakes, NJ) as described previously (52). Staphylococci counts were performed by filtering 100 mL of melted ice samples through Microfil 0.45-μm-pore-size membranes (EMD Millipore, Billerica, MA), which were then placed on Baird-Parker plates (G96, Hardy Diagnostics, Santa Maria, CA). After incubation at 37°C for up to 48 h, growth of black colonies was recorded, and these colonies were further tested for coagulase using CoaguStaph (Z020, Hardy Diagnostics). Samples were analyzed for Salmonella using the FDA-validated quantitative PCR procedure with a detection limit of 0.08 to 0.2 CFU/g (10). Composited samples were used in the initial screening by filtering 5 to 10 100-mL subsamples of each sample through a membrane. The membrane was placed in modified buffered peptone water for enrichment culture at 35 ± 2°C for 18 to 24 h. The data were graphed and analyzed with a one-way nonparametric analysis of variance (ANOVA) using Prism software (GraphPad, La Jolla, CA). Kruskal-Wallis tests were used for the ANOVA and multiple comparisons between and among sampling groups.

Microbial community analysis by 16S and 18S rRNA sequencing

Selected samples were analyzed for their microbial community composition using MiSeq next generation sequencing (Illumina, San Diego, CA). Because of the limitations of resources and the amount of DNA available in the ice, our analysis was limited to three packaged ice samples: two on-site packaged ice samples and one IPIA-compliant ISB ice sample designated as on-site A, on-site B, and ISB, respectively. The on-site A and ISB samples were chosen among the samples with satisfactory microbiological quality that contained sufficient DNA for the sequencing analysis. The on-site B sample was chosen to represent the samples that were not PIQCS compliant for microbiological quality. None of the manufactured ice samples produced sufficient DNA for positive sequencing reads. Melted ice of volumes from 300 mL to several liters were filtered to obtain enough cells for DNA extraction using the PowerWater DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, CA) following the manufacturer's procedure. Purified DNA was subjected to amplification using the bacteria and archaea universal primers 515F (GTGCCAGCMGCCGCGGTAA) and 806R (GGACTACHVGGGTWTCTAAT), which cover the V4 region of the 16S rRNA gene and have been recommended for environmental samples (8). The Eukarya specific primers TAReukF (CCAGCASCYGCGGTAATTCC) and TAReukR (ACTTTCGTTCTTGATYRA) were used to amplify the V4 region of 18S rRNA DNA (21). The sequencing was performed on an Illumina MiSeq sequencer at the Research and Testing Laboratory (Lubbock, TX).

Raw amplicons of 16S and 18S rRNA were processed according to the MiSeq standard operating procedure (2, 32) using MOTHUR v. 1.36.1 (46). The 250-bp paired MiSeq reads were combined after removal of reads shorter than 200 bp or reads containing ambiguous base pairs. The remaining reads were aligned to SILVA v. 119 (45), and sequencing errors were reduced using pre.cluster (24). Chimeric sequences identified by UCHIME (14), sequences not classified at the kingdom level, and mitochondrial and chloroplast sequences were excluded from the data set. For 16S rRNA analysis, the remaining high-quality reads were degapped and used for closed reference operational taxonomic unit (OTU) picking using Greengenes v. 13.5 (11) in QIIME v. 1.9.1 (8), where sequences were clustered at 97% sequence similarity using UCLUST (13) and classified with SortMeRNA (31). After removal of singletons (OTUs with only one sequence), the remaining OTUs were normalized by 16S rRNA gene copy number to allow intrasample comparison. The normalized OTU table was passed to the core diversity workflow using a nonphylogenetic diversity parameter and an e-value of 17,000 sequences per sample. For 18S rRNA analysis, the remaining high-quality reads were degapped and used for closed reference OTU picking using SILVA v. 119 in QIIME v. 1.9.1, where sequences were clustered at 97% sequence similarity and classified with SortMeRNA. The OTU table was passed to the core diversity workflow using a nonphylogenetic diversity parameter and an e-value of 3,065 sequences per sample. Environmental clustering using principal coordinates analysis plots based on a Bray-Curtis distance matrix was performed for β-diversity analysis, and observed OTUs and Chao1 were used for α-diversity evaluation.

Microbiological quality of packaged ice samples

All samples were analyzed for their microbiological quality using TPCs, coliform and E. coli counts, staphylococci counts, and presence or absence of Salmonella. The results are summarized in Table 1, and the distributions of the TPCs and coliform counts are plotted in Figure 1. For the on-site packaged ice products from stores in all four geographic areas in southern California, 11% of the samples had unacceptable TPCs, but none of the ISB and manufactured ice samples had TPCs above the PIQCS compliant limit of 500 CFU/mL. Among the four areas, samples from Orange County had the lowest prevalence of unacceptable TPCs (2 of 30 samples, 7%) based on the PIQCS limit. Orange County samples also had a lower average count of 77 most probable number (MPN)/mL (Fig. 1). In contrast, Los Angeles (LA) County had the highest prevalence of unacceptable TPCs (13%) and a significantly higher average count at 919 MPN/mL (P < 0.05), which is above the PIQCS limit (Table 1 and Fig. 1). The highest TPC was >104 MPN/mL in a sample collected from LA County, and its melted liquid was distinctively turbid. Repeated samples collected from the same store 2 months apart had similar TPCs, indicating a persistent problem (data not shown). Compared with the on-site packaged ice, ISB and manufactured ice had much lower mean TPCs of 3 to 14 MPN/mL compared with 77 to 919 MPN/mL for the on-site packaged ice (Fig. 1). Overall, the on-site packaged ice samples were more likely to be of unsatisfactory quality, and the average TPCs were higher than those for the ISB and manufactured ice samples tested in this study. Samples from LA, San Bernardino and Riverside, and San Diego and Imperial counties had higher TPCs than did at least one group of the non–on-site packaged ice samples (P < 0.05, Fig. 1).

FIGURE 1.

Scatterplots of total heterotroph counts (A) and coliform counts (B) sorted by sources and types of packaged ice products. On-site packaged ice samples were analyzed by geographical area: Los Angeles County (LA), Orange County (OC), San Bernardino and Riverside counties (SB/R), and San Diego and Imperial counties (SD/I). These samples were compared with ice samples packaged by the manufacturer and those obtained from the in-store bagger (ISB), which were both IPIA compliant (IPIA) and noncompliant (non-IPIA). Samples with a zero value were plotted as 0.1 on the log scale for convenience of comparison. Horizontal bars indicate the means of each sample group. Dotted line indicates the PIQCS acceptable limit of ≤500 MPN/mL for heterotrophs. Brackets with asterisks indicate statistical significance: * P < 0.05; *** P < 0.001.

FIGURE 1.

Scatterplots of total heterotroph counts (A) and coliform counts (B) sorted by sources and types of packaged ice products. On-site packaged ice samples were analyzed by geographical area: Los Angeles County (LA), Orange County (OC), San Bernardino and Riverside counties (SB/R), and San Diego and Imperial counties (SD/I). These samples were compared with ice samples packaged by the manufacturer and those obtained from the in-store bagger (ISB), which were both IPIA compliant (IPIA) and noncompliant (non-IPIA). Samples with a zero value were plotted as 0.1 on the log scale for convenience of comparison. Horizontal bars indicate the means of each sample group. Dotted line indicates the PIQCS acceptable limit of ≤500 MPN/mL for heterotrophs. Brackets with asterisks indicate statistical significance: * P < 0.05; *** P < 0.001.

Close modal

Of the 30 samples tested from each area, 3, 3, and 6 samples from LA, Orange, and San Diego and Imperial counties, respectively, were positive for coliforms, as indicated by the fluorescent colonies on the MI plates (Table 1). On-site packaged ice samples from San Bernardino and Riverside counties and manufactured and ISB ice samples had no detectable coliforms. None of the samples tested were positive for E. coli, indicating a low likelihood of fecal contamination in the samples analyzed. The one-way nonparametric ANOVA revealed no significant difference in coliform counts among the seven groups (P > 0.05), probably because of the smaller number of samples in the ISB and manufactured ice groups. However, when analyzing the on-site packaged ice samples versus the combined non–on-site packaged ice samples (ISB and manufactured combined), samples from San Diego and Imperial counties had significantly higher coliform counts than did samples from San Bernardino and Riverside counties (P = 0.035) or the non–on-site packaged samples (P = 0.023).

When applying the PIQCS microbiological criteria for TPCs of ≤500 MPN/mL and zero coliforms and E. coli per 100 mL, the on-site packaged ice had an overall acceptable rate of 81%; the samples with the lowest counts originated in convenience stores in San Bernardino and Riverside counties (90%), followed by Orange County (83%), LA County (80%), and San Diego and Imperial counties (70%) (Table 1). All manufactured and ISB ice samples (100%) were acceptable. Our results confirm the high microbiological quality of manufactured ice, as previously reported from other regions (38, 39). We also determined that the ISBs (whether IPIA compliant or not) produced ice products of satisfactory microbiological quality, possibly because of lack of human handling. When comparing results of TPC analyses of on-site packaged ice from different studies, 11% of our on-site samples had unacceptable TPCs (Table 1), whereas the unacceptable TPC rate was 6% in Georgia (38), 36% in Iowa (39), and 11% in the United Kingdom (43). The percentage of samples positive for coliforms was 10% for our on-site packaged ice samples (Table 1), 37% in both Georgia and Greece (18, 38), 13.5% in Florida (47), 4.5% in Iowa (39), and 9% in the United Kingdom and Hong Kong (20, 43). The variations among these studies may be attributed to the sources of water, the nature of the samples, and the testing methods. Nevertheless, in all studies the microbiological quality of the on-site packaged ice was lower than that of the manufactured ice.

Additional tests were performed to determine the presence of Salmonella and staphylococci (Table 1). None of the 156 samples were positive for Salmonella. This result was similar to those found previously; Salmonella was found in only 1 of 275 ice samples in Georgia (38) and was not found in other studies (20, 39, 43, 47). In Greece, Salmonella was found in 4% of the ice used to cool drinks, which may be a severe public health concern (18).

Staphylococci are part of the human skin microbiota and are routinely used as indicators of contamination attributed to food handlers (17, 34). Staphylococci were detected (based on the growth on Baird Parker agar) in 34% of the on-site packaged ice samples but not in the ISB or manufactured ice samples. The on-site packaging process probably introduced these bacteria into the ice through human contact. The highest staphylococcal count was 528 CFU/100 mL, which was found in one of the LA County samples. The presence of these bacteria was detected in a second sample collected 2 months later, which indicated the persistence of the problem. Among the samples from San Diego and Imperial counties, which had the poorest quality based on TPCs and counts of coliforms and E. coli (Table 1), the number of samples positive for staphylococci was relatively low (8 of 30) compared with 10 or more positive samples from the other counties. The staphylococcal isolates were further tested for coagulase activity as an indication of the enterotoxigenic foodborne pathogen Staphylococcus aureus. Four of 41 staphylococcal isolates from these samples had weak coagulase activity, suggesting a positive result for S. aureus. Staphylococci were isolated from ice samples in Iowa (39) and from 144 of 320 rural drinking water samples evaluated at Oregon State University (34); among these 144 positive samples, 20 samples contained S. aureus.

Microbial community analysis overview

To examine the microbial diversity and abundance in packaged ice products, next generation sequencing analysis was performed using 16S and 18S rRNA gene targeted sequencing to determine the prokaryotic (bacteria and archaea) and eukaryotic microbial communities, respectively. Two on-site packaged and one IPIA-compliant ISB packaged ice products were sampled: on-site A, on-site B, and ISB, respectively. The on-site A and ISB samples were chosen among the samples with satisfactory microbiological quality from which sufficient DNA could be obtained for the sequencing analysis. The on-site B sample was chosen to represent the samples that did not meet the PIQCS standards. None of the manufactured ice produced sufficient DNA for positive sequencing reads. The diversity of microbial species is presented as the total OTUs. Each OTU represents a unique 16S or 18S rRNA taxonomic identity. All three samples generated at least 17,000 sequences from the 16S rRNA analysis. The two on-site packaged ice products had 194 and 136 OTUs, respectively. In contrast, the ISB packaged ice had much lower microbial diversity, with only 29 OTUs, most likely because its production occurred in a closed environment without human handling. Compared with the 16S rRNA data, the 18S rRNA analysis produced 10 times fewer sequencing reads for all three samples and revealed lower eukaryote diversity, at 2, 15, and 6 OTUs for on-site A, on-site B, and ISB samples, respectively. These results suggest that bacteria dominate the microbial communities in the ice products analyzed in this study. However, the abundance of the bacteria and eukaryotes cannot be directly compared because of the use of different sequencing primer sets.

Compared with the levels of viable counts, the number of OTUs is not always correlated with the TPCs. For example, the TPC for the on-site A sample was 128 MPN/mL, and the TPC for the on-site B sample was 2.4 × 104 MPN/mL. Despite its lower TPC, the on-site A sample was more diverse (194 OTUs) than the on-site B sample (136 OTUs). Although sequencing analysis cannot discriminate among cells that are viable, nonculturable, or dead unless some additional effort is used to selectively amplify DNA from viable cells (53), this further discrimination was beyond the scope of this study.

Principal coordinates analysis plots were used to visualize the correlations among the sample sets based on Bray-Curtis dissimilarity, observed OTUs, and Chao1 (Fig. 2). The 16S rRNA profiles of the two on-site packaged ice samples are similar to each other, as indicated by the close locations on the plots with either the α- or β-analysis. The 16S rRNA profile of ISB packaged ice is distinctive from those of the on-site packaged ice products, as indicated by the distal location of the coordinates on both plots. The 18S rRNA eukaryote profiles of all three samples are distinct from the bacterial and archaeal profiles as expected. A Venn diagram was created to compare the unique and shared prokaryote OTUs among the three samples. The two on-site packaged ice samples shared 41 OTUs, whereas the ISB sample shared fewer than 5 OTUs with either of the two on-site samples (Fig. 2C). Both the principal coordinates analysis plots and the Venn diagram agreed that the two on-site packaged ice samples have similar 16S rRNA profiles, which are different from the profile of the ISB ice sample.

FIGURE 2.

Principal coordinates analysis plots of microbial communities based on Bray-Curtis dissimilarity analysis (A and B) and a Venn diagram of bacterial OTUs shared among the three samples (C). Open and closed symbols represent 16S and 18S rRNA sequencing results, respectively. (A) Principal component 1 (PC1), which explains 20.72% variance, is plotted against PC2 (20.12% variance explained). (B) PC1 is plotted against PC3, which explains 20.06% variance.

FIGURE 2.

Principal coordinates analysis plots of microbial communities based on Bray-Curtis dissimilarity analysis (A and B) and a Venn diagram of bacterial OTUs shared among the three samples (C). Open and closed symbols represent 16S and 18S rRNA sequencing results, respectively. (A) Principal component 1 (PC1), which explains 20.72% variance, is plotted against PC2 (20.12% variance explained). (B) PC1 is plotted against PC3, which explains 20.06% variance.

Close modal

16S rRNA sequence analysis of bacterial and archaeal communities in ice

The identities of bacteria and archaea were determined by mapping the sequences to the Greengenes v. 13.5 16S rRNA database. The abundance data were calculated based on the sequence counts of each OTU divided by the copy number of each single species. Proteobacteria was the dominant phylum for all three samples, at approximately 90% for the two on-site packaged ice samples and almost 100% for the ISB packaged ice sample analyzed. Other than Proteobacteria, the on-site A and B samples also included OTUs in the phyla Actinobacteria, Bacteroidetes, Fermicutes, and Verrucomicrobia, which are known to abundantly inhabit soil, environments, and guts of humans and animals (28, 44). None of the ice products had any Archaea, in corroboration with other reports that Archaea are rare (at low levels) in drinking water systems (19).

Within the Proteobacteria, Betaproteobacteria dominated the on-site A sample (72%) and Alphaproteobacteria dominated the on-site B and ISB samples at 63.55 and 99.97%, respectively (Fig. 3). Proteobacteria are the primary inhabitants of the ground water system (29). Betaproteobacteria were reported as 80 to 98% of the bacteria in bottled mineral water produced in Denmark and Norway as determined by fluorescence in situ hybridization and 16S rRNA sequence analysis (37). The Gammaproteobacteria, including coliforms and many others, had low prevalence in all three samples analyzed, at 5.0, 0.6, and 0.006% of the on-site A, on-site B, and ISB samples, respectively (Fig. 3).

FIGURE 3.

Relative abundance of bacterial communities at the class level for the two on-site packaged ice samples (site A and site B) and the ISB packaged ice sample analyzed using 16S rRNA targeted sequencing. Individual OTUs with >1% abundance were plotted; OTUs with <1% abundance were included in the “Other” category.

FIGURE 3.

Relative abundance of bacterial communities at the class level for the two on-site packaged ice samples (site A and site B) and the ISB packaged ice sample analyzed using 16S rRNA targeted sequencing. Individual OTUs with >1% abundance were plotted; OTUs with <1% abundance were included in the “Other” category.

Close modal

The relative abundance of the major genera or families (when the genera were not identified) are summarized in Table 2. Members of the Caulobacteraceae (51%) and Sphingobium (43%) were the majority of bacteria in the ISB packaged ice analyzed in this study. These two taxa were also present in the two on-site packaged ice samples, at <2% relative abundance, implicating their common presence in the packaged ice samples. Both Caulobacteraceae and Sphingobium are Alphaproteobacteria and can be abundant in chlorinated drinking water, bottled mineral water, ground water, and glacial ice (3, 19, 29, 37, 48). These microorganisms can also be abundant in aquatic environments as part of biofilms (1, 23, 54). The presence of low levels of both Caulobacteraceae and Sphingobium in our ISB and on-site packaged ice samples suggests the presence of biofilms in the ice machine and/or the system from which the water originated. No health concerns have been associated with these microorganisms.

TABLE 2.

Bacterial identification and relative abundance in the packaged ice samples analyzed by 16S rRNA sequencing

Bacterial identification and relative abundance in the packaged ice samples analyzed by 16S rRNA sequencing
Bacterial identification and relative abundance in the packaged ice samples analyzed by 16S rRNA sequencing

Both of the on-site packaged ice samples analyzed in this study contain a diverse collection of bacterial OTUs (Figs. 2C and 3 and Table 2). The on-site A sample, which was compliant with the PIQCS, was dominated by methylotrophs such as Methylophilaceae, Methylibium, and Comamonadaceae at 31, 30, and 5% relative abundance, respectively (Table 2). The Methylophilaceae is prevalent in surface water and terrestrial and marine environments and plays a role in carbon cycling (4). Members of the Comamonadaceae, including Methylibium, have been reported as the primary bacteria in activated waste water sludge (30) and gasoline-contaminated aquifers (42, 49). Therefore, these methylotrophs could have been enriched and thus survived the water treatment process.

The on-site B ice sample, which was not PIQCS compliant, was dominated by photosynthetic purple nonsulfur Alphaproteobacteria in the family of Rhodospirillaceae at 53% relative abundance (including the genus Phaeospirillum) (Table 2). Twelve unique OTUs were identified for this family, indicating its abundance and diversity in the on-site B ice sample. Members of the Rhodospirillaceae can be found in anaerobic aquatic environments such as mud and stagnant water. These bacteria have been found in ground and well water at up to 10% abundance (29). The source water for the on-site B ice sample probably was either untreated or polluted with ground and/or surface water, resulting in contamination with Rhodospirillaceae. The water also may have been subjected to abusive conditions that encouraged the growth of Rhodospirillaceae and Comamonadaceae to high levels.

Some potential or opportunistic pathogens were found at low abundance in the two on-site packaged ice samples, suggesting a potential public health threat. Among the significant taxa, the Firmicutes, Staphylococcus, and Burkholderia were present in the on-site B sample at low abundance (≤0.1%) but were not found in the on-site A or ISB packaged ice. The results are in agreement with our staphylococci selective plating data (data not shown). The on-site A sample contained Mycobacterium at 4.6% (Table 2). Legionella, Rickettsiales, and Pseudomonas were found in both on-site A and B packaged ice samples at <0.1% abundance (data not shown). Mycobacterium and Legionella cause some respiratory infections and were found in monochloramine-treated drinking water and chlorinated water samples, respectively (19). Enterobacteriaceae, including coliforms and pathogens such as Salmonella and E. coli, were absent from all three samples, indicating that fecal contamination was unlikely in the ice samples analyzed. The lack of Enterobacteriaceae was also reported in studies of drinking water (29).

18S rRNA eukaryote community analysis

The compositions of eukaryotic microorganisms were analyzed using 18S rRNA gene sequencing for the two on-site packaged ice samples and one ISB ice sample. The relative abundance of each OTU is listed in Table 3. The majority of the eukaryotic microorganisms identified were fungi, although some unicellular algae and protists were also identified. ISB packaged ice had one dominant yeast species, Bensingtonia yamatoana (98%), and five other minor species of fungi, algae, and protists. B. yamatoana is a basidiomycetous budding yeast that has been found on the surfaces of plants (41) and is phylogenetically related to Rhodotorula arctica found in arctic soil (55). However, its prevalence in water and other sources has not been reported. The ISB ice also contained other yeasts such as Rhodotorula, Microbotryum violaceum, and Sporobolomyces, which are known to be prevalent in oligotrophic water and deep ground water and can originate from the microbial effluents of plants (40).

TABLE 3.

Eukaryotic OTUs and relative abundance in the packaged ice samples analyzed by 18S rRNA sequencing

Eukaryotic OTUs and relative abundance in the packaged ice samples analyzed by 18S rRNA sequencing
Eukaryotic OTUs and relative abundance in the packaged ice samples analyzed by 18S rRNA sequencing

The on-site A sample contained only two OTUs of the same yeast species, Saccharomyces cerevisiae (Table 3). S. cerevisiae can be found in human environments because of its common use in brewing and baking and in its natural habitat on the surface of fruits and trees (35). The on-site B ice, which was not PIQCS compliant, contained diverse eukaryotic microbes, as expected. The dominant groups were the uncultured Tremellales yeasts (48%) and the two algal OTUs, a chrysophytic alga (46%) and the Spumella-like flagellate 49D3 (3%) (Table 3). Tremellales is a group of basidiomecetous yeasts that includes opportunistic pathogens such as Cryptococcus (36). A separate OTU (Cryptococcus carnescens) was found in the same sample at 0.1% prevalence (Table 3) and has occasionally been associated with opportunistic infections (15, 50). Cryptococcus has been found in indoor dust (22). Chrysophytes and Spumella are major groups of nanoflagellates in fresh waters and soil of various geographical regions (5). Additional evidence of water and/or soil contamination in the on-site B sample was the presence of three protozoan species, Vermamoeba vermiformis, Protostelium nocturnum, and Cercozoa (Table 3). V. vermiformis is a free-living amoeba prevalent in the environment and in treated drinking water, especially in hot water systems, probably because of its resistant cysts (7, 16). These amoebae could be fed by biofilms and waterborne pathogens such as Legionella (33). In our study, several Legionella OTUs were found in both the on-site A and on-site B ice samples, indicating a potential interaction between the amoebae and the potentially pathogenic bacteria. The on-site B sample also contained genes from sesame plants and/or seeds and from tree shrews, which may further indicate that the water was contaminated with material from animals, plants, and/or foods. The analysis of the eukaryotic microbial community again supports the hypothesis that the on-site B ice may have been made from surface or well water, may have been cross-contaminated with environmental material, and/or may have been abused for an extended time.

In conclusion, a higher percentage of the on-site packaged ice samples were of unacceptable microbiological quality (19%), and these samples had higher TPCs and counts of coliforms and staphylococci. In contrast, 100% of the manufactured and ISB ice samples were within the acceptable range for microbiological quality, had lower TPCs, and produced zero growth of coliforms and staphylococci. None of the 156 samples analyzed were positive for E. coli or Salmonella, indicating the lack of enteric pathogens in these samples. Overall, the on-site packaged ice samples were of poorer microbiological quality than the manufactured and ISB ice. This difference for the on-site packaged samples may be attributed to contamination from the environment and/or food handlers and lack of quality controls based on the PIQCS or other processing standards.

The microbial community analysis of these ice products revealed that bacteria were more abundant and diverse than eukaryotic microorganisms. Archaea were not detected in the three ice samples sequenced in this study. The bacterial communities were more diverse in the on-site packaged ice samples than in the ISB packaged ice. Proteobacteria, especially Alpha- and Betaproteobacteria, was the dominant bacterial group in all three samples. Both Alpha- and Betaproteobacteria are widely distributed in the environment, soil, and water. Most of the species present in the PIQCS-compliant ice samples (on-site A and ISB) were oligotrophic bacteria and yeasts that may have survived water treatments and/or been persistent as biofilms in the systems, as has been reported for drinking water (19, 29). The ice sample that was not PIQCS compliant (on-site B) contained a large number of photosynthetic bacteria and some algae and protists, suggesting that the water source for this ice was cross-contaminated by surface water and/or stored under abusive conditions. Several opportunistic or potential pathogens were found at low abundance in the on-site packaged ice but not in the ISB packaged ice, raising food safety concerns for on-site packaged ice. The 16S and 18S targeted sequencing analysis revealed minute details of the microbial communities in our samples and provided a potential way to trace the source of contamination in packaged ice. These findings support our hypothesis that ice packaged at a retail site has higher bacterial diversity and a greater risk of pathogen contamination than ice packaged by a machine or the manufacturer. To our knowledge, this is the first study of packaged ice products in which the microbial communities were identified through next generation sequencing analysis.

The project was supported by the IPIA. Part of the research was carried out at the Jet Propulsion Laboratory (California Institute of Technology, Pasadena) under a contract with the National Aeronautics and Space Administration. M. Bashir was financed by the FWF Austrian Science Fund, the DK-MOLIN program, and the Bank Austria Visiting Scientist Program. We thank Nancy Lieu (La Verne Metropolitan Water District, La Verne, CA) and Mahjabeen Ahmed (California State Polytechnic University, Pomona) for technical help. We also thank Sonya Liu for editorial review.

1
Abraham,
W.-R.,
M.
Rohde,
and
A.
Bennasar.
2014
.
The family Caulobacteraceae,
p
.
179
205
.
In
E.
Rosenberg,
E. F.
DeLong,
S.
Lory,
E.
Stackebrandt,
and
F.
Thompson
(
ed.
),
The prokaryotes: Alphaproteobacteria and Betaproteobacteria
.
Springer
,
Berlin
.
2
Anonymous
.
2015
.
MiSeq SOP
.
Available at: https://www.mothur.org/wiki/MiSeq_SOP. Accessed 1 November 2015
.
3
Balkwill,
D. L.,
J. K.
Fredrickson,
and
M. F.
Romine.
2006
.
Sphingomonas and related genera,
p
.
605
629
.
In
M.
Dworkin,
S.
Falkow,
E.
Rosenberg,
K.-H.
Schleifer,
and
E.
Stackebrandt
(
ed.
),
The prokaryotes, vol. 7
.
Proteobacteria: delta, epsilon subclass
.
Springer, New York.
4
Beck,
D. A. C.,
T. L.
McTaggart,
U.
Setboonsarng,
A.
Vorobev,
M. G.
Kalyuzhnaya,
N.
Ivanova,
L.
Goodwin,
T.
Woyke,
M. E.
Lidstrom,
and
L.
Chistoserdova.
2014
.
The expanded diversity of Methylophilaceae from Lake Washington through cultivation and genomic sequencing of novel ecotypes
.
PLoS ONE
9
:
e102458
.
5
Boenigk,
J.,
K.
Pfandl,
P.
Stadler,
and
A.
Chatzinotas.
2005
.
High diversity of the “Spumella-like” flagellates: an investigation based on the SSU rRNA gene sequences of isolates from habitats located in six different geographic regions
.
Environ. Microbiol
.
7
:
685
697
.
6
Burgess,
C. M.,
A.
Gianotti,
N.
Gruzdev,
J.
Holah,
S.
Knøchel,
A.
Lehner,
E.
Margas,
S. S.
Esser,
S.
Sela,
and
O.
Tresse.
2016
.
The response of foodborne pathogens to osmotic and desiccation stresses in the food chain
.
Int. J. Food Microbiol
.
221
:
37
53
.
7
Buse,
H. Y.,
J.
Lu,
I. T.
Struewing,
and
N. J.
Ashbolt.
2013
.
Eukaryotic diversity in premise drinking water using 18S rDNA sequencing: implications for health risks
.
Environ. Sci. Pollut. Res
.
20
:
6351
6366
.
8
Caporaso,
J. G.,
J.
Kuczynski,
J.
Stombaugh,
K.
Bittinger,
F. D.
Bushman,
E. K.
Costello,
N.
Fierer,
A. G.
Peña,
J. K.
Goodrich,
J. I.
Gordon,
G. A.
Huttley,
S. T.
Kelley,
D.
Knights,
J. E.
Koenig,
R. E.
Ley,
C. A.
Lozupone,
D.
McDonald,
B. D.
Muegge,
M.
Pirrung,
J.
Reeder,
J. R.
Sevinsky,
P. J.
Turnbaugh,
W. A.
Walters,
J.
Widmann,
T.
Yatsunenko,
J.
Zaneveld,
and
R.
Knight.
2010
.
QIIME allows analysis of high-throughput community sequencing data
.
Nat. Methods
7
:
335
336
.
9
Centers for Disease Control
.
1987
.
Epidemiologic notes and reports outbreak of viral gastroenteritis—Pennsylvania and Delaware. Morb. Mortal. Wkly. Rep. Surveill
.
Summ
.
36
:
709
711
.
10
Cheng,
C.-M.,
K. T.
Van,
W.
Lin,
and
R. M.
Ruby.
2009
.
Interlaboratory validation of a real-time PCR 24-hour rapid method for detection of Salmonella in foods
.
J. Food Prot
.
72
:
945
951
.
11
DeSantis,
T. Z.,
P.
Hugenholtz,
N.
Larsen,
M.
Rojas,
E. L.
Brodie,
K.
Keller,
T.
Huber,
D.
Dalevi,
P.
Hu,
and
G. L.
Andersen.
2006
.
Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB
.
Appl. Environ. Microbiol
.
72
:
5069
5072
.
12
de Souza Sant'Ana,
A.
2012
.
Introduction to the special issue: Salmonella in foods: evolution, strategies and challenges
.
Food Res. Int
.
45
:
451
454
.
13
Edgar,
R. C.
2010
.
Search and clustering orders of magnitude faster than BLAST
.
Bioinformatics
26
:
2460
2461
.
14
Edgar,
R. C.,
B. J.
Haas,
J. C.
Clemente,
C.
Quince,
and
R.
Knight.
2011
.
UCHIME improves sensitivity and speed of chimera detection
.
Bioinformatics
27
:
2194
2200
.
15
Ferreira-Paim,
K.,
T. B.
Ferreira,
L.
Andrade-Silva,
D. J.
Mora,
D. J.
Springer,
J.
Heitman,
F. M.
Fonseca,
D.
Matos,
M. S. C.
Melhem,
and
M. L.
Silva-Vergara.
2014
.
Phylogenetic analysis of phenotypically characterized Cryptococcus laurentii isolates reveals high frequency of cryptic species
.
PLoS ONE
9
:
e108633
.
16
Fouque,
E.,
M.-C.
Trouilhé,
V.
Thomas,
P.
Humeau,
and
Y.
Héchard.
2014
.
Encystment of Vermamoeba (Hartmannella) vermiformis: effects of environmental conditions and cell concentration
.
Exp. Parasitol
.
145
:
S62
S68
.
17
Gabutti,
G.,
A.
De Donno,
F.
Bagordo,
and
M. T.
Montagna.
2000
.
Comparative survival of faecal and human contaminants and use of Staphylococcus aureus as an effective indicator of human pollution
.
Mar. Pollut. Bull
.
40
:
697
700
.
18
Gerokomou,
V.,
C.
Voidarou,
A.
Vatopoulos,
E.
Velonakis,
G.
Rozos,
A.
Alexopoulos,
S.
Plessas,
E.
Stavropoulou,
E.
Bezirtzoglou,
P. G.
Demerzis,
and
K.
Akrida-Demertzi.
2011
.
Physical, chemical and microbiological quality of ice used to cool drinks and foods in Greece and its public health implications
.
Anaerobe
17
:
351
353
.
19
Gomez-Alvarez,
V.,
R. P.
Revetta,
and
J. W.
Santo Domingo
.
2012
.
Metagenomic analyses of drinking water receiving different disinfection treatments
.
Appl. Environ. Microbiol
.
78
:
6095
6102
.
20
Government of Hong Kong
.
2005
.
The microbiological quality of edible ice from ice manufacturing plants and retail business in Hong Kong. Document 21
.
Government of the Hong Kong Special Administrative Region
,
Hong Kong
.
21
Guillou,
L.,
D.
Bachar,
S.
Audic,
D.
Bass,
C.
Berney,
L.
Bittner,
C.
Boutte,
G.
Burgaud,
C.
de Vargas,
J.
Decelle,
J.
del Campo,
J. R.
Dolan,
M.
Dunthorn,
B.
Edvardsen,
M.
Holzmann,
W. H. C. F.
Kooistra,
E.
Lara,
N.
Le Bescot,
R.
Logares,
F.
Mahé,
R.
Massana,
M.
Montresor,
R.
Morard,
F.
Not,
J.
Pawlowski,
I.
Probert,
A.-L.
Sauvadet,
R.
Siano,
T.
Stoeck,
D.
Vaulot,
P.
Zimmermann,
and
R.
Christen.
2013
.
The protist ribosomal reference database (PR2): a catalog of unicellular eukaryote small sub-unit rRNA sequences with curated taxonomy
.
Nucleic Acids Res
.
41
:
D597
D604
.
22
Hanson,
B.,
Y.
Zhou,
E. J.
Bautista,
B.
Urch,
M.
Speck,
F.
Silverman,
M.
Muilenberg,
W.
Phipatanakul,
G.
Weinstock,
E.
Sodergren,
D. R.
Gold,
and
J. E.
Sordillo.
2016
.
Characterization of the bacterial and fungal microbiome in indoor dust and outdoor air samples: a pilot study. Environ. Sci. Process
.
Impacts
18
:
713
724
.
23
Hong,
P.-Y.,
C.
Hwang,
F.
Ling,
G. L.
Andersen,
M. W.
LeChevallier,
and
W.-T.
Liu.
2010
.
Pyrosequencing analysis of bacterial biofilm communities in water meters of a drinking water distribution system
.
Appl. Environ. Microbiol
.
76
:
5631
5635
.
24
Huse,
S. M.,
D. M.
Welch,
H. G.
Morrison,
and
M. L.
Sogin.
2010
.
Ironing out the wrinkles in the rare biosphere through improved OTU clustering
.
Environ. Microbiol
.
12
:
1889
1898
.
25
International Packaged Ice Association
.
2013
.
The PIQCS manual (packaged ice quality control standards)
.
International Packaged Ice Association
,
Tampa, FL
.
26
International Packaged Ice Association
.
2016
.
About the packaged ice industry
.
Available at: http://www.safeice.org/. Accessed 15 March 2016
.
27
International Packaged Ice Association
.
17 May 2016
.
Personal communication
.
28
Jovel,
J.,
J.
Patterson,
W.
Wang,
N.
Hotte,
S.
O'Keefe,
T.
Mitchel,
T.
Perry,
D.
Kao,
A. L.
Mason,
K. L.
Madsen,
and
G. K. S.
Wong.
2016
.
Characterization of the gut microbiome using 16S or shotgun metagenomics
.
Front. Microbiol
.
7
:
459
. doi:.
29
Karwautz,
C.,
and
T.
Lueders.
2014
.
Impact of hydraulic well restoration on native bacterial communities in drinking water wells
.
Microbes Environ
.
29
:
363
369
.
30
Khan,
S. T.,
Y.
Horiba,
M.
Yamamoto,
and
A.
Hiraishi.
2002
.
Members of the family Comamonadaceae as primary poly(3-hydroxybutyrate-co-3-hydroxyvalerate)-degrading denitrifiers in activated sludge as revealed by a polyphasic approach
.
Appl. Environ. Microbiol
.
68
:
3206
3214
.
31
Kopylova,
E.,
L.
Noe,
and
H.
Touzet.
2012
.
SortMeRNA: fast and accurate filtering of ribosomal RNAs in metatranscriptomic data
.
Bioinformatics
28
:
3211
3217
.
32
Kozich,
J. J.,
S. L.
Westcott,
N. T.
Baxter,
S. K.
Highlander,
and
P. D.
Schloss.
2013
.
Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform
.
Appl. Environ. Microbiol
.
79
:
5112
5120
.
33
Lau,
H. Y.,
and
N. J.
Ashbolt.
2009
.
The role of biofilms and protozoa in Legionella pathogenesis: implications for drinking water
.
J. Appl. Microbiol
.
107
:
368
378
.
34
LeChevallier,
M. W.,
and
R. J.
Seidler.
1980
.
Staphylococcus aureus in rural drinking water
.
Appl. Environ. Microbiol
.
30
:
739
742
.
35
Liti,
G.
2015
.
The fascinating and secret wild life of the budding yeast S. cerevisiae
.
eLife
4
:
e05835
.
36
Liu,
X. Z.,
Q. M.
Wang,
B.
Theelen,
M.
Groenewald,
F. Y.
Bai,
and
T.
Boekhout.
2015
.
Phylogeny of tremellomycetous yeasts and related dimorphic and filamentous basidiomycetes reconstructed from multiple gene sequence analyses
.
Stud. Mycol
.
81
:
1
26
.
37
Loy,
A.,
W.
Beisker,
and
H.
Meier.
2005
.
Diversity of bacteria growing in natural mineral water after bottling
.
Appl. Environ. Microbiol
.
71
:
3624
3632
.
38
Mako,
S. L.,
M. A.
Harrison,
V.
Sharma,
and
F.
Kong.
2014
.
Microbiological quality of packaged ice from various sources in Georgia
.
J. Food Prot
.
77
:
1546
1553
.
39
Moyer,
N. P.,
G. M.
Breuer,
N. H.
Hall,
J. L.
Kempf,
L. A.
Friell,
G. W.
Ronald,
and
J. W. J.
Hausler.
1993
.
Quality of packaged ice purchased at retail establishments in Iowa
.
J. Food Prot
.
56
:
426
431
.
40
Nagahama,
T.
2006
.
Yeast biodiversity in freshwater, marine and deep-sea environments
,
p
.
241
262
.
In
C. A.
Rosa
and
G.
Peter
(
ed.
),
Biodiversity and ecophysiology of yeasts
.
Springer
,
New York
.
41
Nakase,
T.
2000
.
Expanding world of ballistosporous yeasts: distribution in the phyllosphere, systematics and phylogeny
.
J. Gen. Appl. Microbiol
.
46
:
189
216
.
42
Nakatsu,
C. H.,
K.
Hristova,
S.
Hanada,
X.-Y.
Meng,
J. R.
Hanson,
K. M.
Scow,
and
Y.
Kamagata.
2006
.
Methylibium petroleiphilum gen. nov., sp. nov., a novel methyl tert-butyl ether-degrading methylotroph of the Betaproteobacteria
.
Int. J. Syst. Evol. Microbiol
.
56
:
983
989
.
43
Nichols,
G.,
I.
Gillespie,
and
J.
de Louvois.
2000
.
The microbiological quality of ice used to cool drinks and ready-to-eat food from retail and catering premises in the United Kingdom
.
J. Food Prot
.
63
:
78
82
.
44
Penton,
C.,
V.
Gupta,
J.
Yu,
and
J.
Tiedje.
2016
.
Size matters: assessing optimum soil sample size for fungal and bacterial community structure analyses using high throughput sequencing of rRNA gene amplicons
.
Front. Microbiol
.
7
:
824
. doi:.
45
Quast,
C.,
E.
Pruesse,
P.
Yilmaz,
J.
Gerken,
T.
Schweer,
P.
Yarza,
J.
Peplies,
and
F. O.
Glockner.
2012
.
The SILVA ribosomal RNA gene database project: improved data processing and Web-based tools
.
Nucleic Acids Res
.
41
:
D590
D596
.
46
Schloss,
P. D.,
S. L.
Westcott,
T.
Ryabin,
J. R.
Hall,
M.
Hartmann,
E. B.
Hollister,
R. A.
Lesniewski,
B. B.
Oakley,
D. H.
Parks,
C. J.
Robinson,
J. W.
Sahl,
B.
Stres,
G. G.
Thallinger,
D. J.
Van Horn,
and
C. F.
Weber.
2009
.
Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities
.
Appl. Environ. Microbiol
.
75
:
7537
7541
.
47
Schmidt,
R. H.,
and
G. E.
Rodrick.
1999
.
Microbial, physical, and chemical quality of packaged ice in Florida
.
J. Food Prot
.
62
:
526
531
.
48
Simon,
C.,
A.
Wiezer,
A. W.
Strittmatter,
and
R.
Daniel.
2009
.
Phylogenetic diversity and metabolic potential revealed in a glacier ice metagenome
.
Appl. Environ. Microbiol
.
75
:
7519
7526
.
49
Szabó,
Z.,
P.
Gyula,
H.
Robotka,
E.
Bató,
B.
Gálik,
P.
Pach,
P.
Pekker,
I.
Papp,
and
Z.
Bihari.
2015
.
Draft genome sequence of Methylibium sp. strain T29, a novel fuel oxygenate-degrading bacterial isolate from Hungary. Stand
.
Genomic Sci
.
10
:
39
. doi:.
50
Takashima,
M.,
T.
Sugita,
T.
Shinoda,
and
T.
Nakase.
2003
.
Three new combinations from the Cryptococcus laurentii complex: Cryptococcus aureus, Cryptococcus carnescens and Cryptococcus peneaus
.
Int. J. Syst. Evol. Microbiol
.
53
:
1187
1194
.
51
Townsend,
D. E.,
and
A.
Naqui.
1998
.
Comparison of SimPlate total plate count test with plate count agar method for detection and quantitation of bacteria in food
.
J. AOAC Int
.
81
:
563
569
.
52
U.S. Environmental Protection Agency
.
2002
.
Method 1604: total coliforms and Escherichia coli in water by membrane filtration using a simultaneous detection technique (MI medium). EPA-821-R-02-024
.
Office of Water, U.S. Environmental Protection Agency
,
Washington, DC
.
53
Vaishampayan,
P.,
A. J.
Probst,
M. T.
La Duc,
E.
Bargoma,
J. N.
Benardini,
G. L.
Andersen,
and
K.
Venkateswaran.
2013
.
New perspectives on viable microbial communities in low-biomass cleanroom environments
.
ISME J
.
7
:
312
324
.
54
Vaz-Moreira,
I.,
O. C.
Nunes,
and
C. M.
Manaia.
2011
.
Diversity and antibiotic resistance patterns of Sphingomonadaceae isolates from drinking water
.
Appl. Environ. Microbiol
.
77
:
5697
5706
.
55
Vishniac,
H. S.,
and
M.
Takashima.
2010
.
Rhodotorula arctica sp. nov., a basidiomycetous yeast from arctic soil
.
Int. J. Syst. Evol. Microbiol
.
60
:
1215
1218
.