By shedding light on variation in time as well as in space, long-term biogeographic studies can help us define organisms’ distribution patterns and understand their underlying drivers. Here we examine distributions of Pseudomonas in and around 15 human homes, focusing on the P. putida and P. fluorescens species groups. We describe recovery from 10,941 samples collected during up to 8 visits per home, occurring on average 2.6 times per year. We collected a mean of 141 samples per visit, from sites in most rooms of the house, from the surrounding yards, and from human and pet occupants. We recovered Pseudomonas in 9.7% of samples, with the majority of isolates being from the P. putida and P. fluorescens species groups (approximately 62% and 23% of Pseudomonas samples recovered respectively). Although representatives of both groups were recovered from every season, every house, and every type of environment sampled, recovery was highly variable across houses and samplings. Whereas recovery of P. putida group was higher in summer and fall than in winter and spring, P. fluorescens group isolates were most often recovered in spring. P. putida group recovery from soils was substantially higher than its recovery from all other environment types, while higher P. fluorescens group recovery from soils than from other sites was much less pronounced. Both species groups were recovered from skin and upper respiratory tract samples from healthy humans and pets, although this occurred infrequently. This study indicates that even species that are able to survive under a broad range of conditions can be rare and variable in their distributions in space and in time. For such groups, determining patterns and causes of stochastic and seasonal variability may be more important for understanding the processes driving their biogeography than the identity of the types of environments in which they can be found.
Citation: Remold SK, Purdy-Gibson ME, France MT, Hundley TC (2015) Pseudomonas putida and Pseudomonas fluorescens Species Group Recovery from Human Homes Varies Seasonally and by Environment. PLoS ONE 10(5): e0127704. https://doi.org/10.1371/journal.pone.0127704
Academic Editor: George-John Nychas, Agricultural University of Athens, GREECE
Received: February 2, 2015; Accepted: April 17, 2015; Published: May 29, 2015
Copyright: © 2015 Remold et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: Raw data describing the characteristics of the sampling time and location, and whether or not each of the focal Pseudomonas, P. putida group and P. fluorescens group were recovered are available in tab delimited text format in the Dryad Digital Repository (DRYAD), doi:10.5061/dryad.2s361. Partial 16S rDNA sequences of Pseudomonas samples are available as Genbank accessions KP452510–KP453692.
Funding: This work was supported by the National Science Foundation, DEB-0950361 to SKR. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The distributions of many organisms vary in time as well as space. In microbial ecology and biogeography, studies that include the longitudinal sampling required to address temporal variation have found significant seasonal variation in communities of soils, [1, 2], rhizospheres , lakes  oceans , and even in the indoor built environment [6–9]. A better understanding of how an organism’s use of its environment varies across seasons has important implications for understanding the parameters that are most important in the delineation of its habitat and also of the types and patterns of selection pressures it experiences. In addition, the degree to which environmental variability exists in space vs. in time has been predicted to affect the life history strategies that are most favorable [10, 11].
We focus here on two large species groups within the genus Pseudomonas, the P. putida and P. fluorescens groups . Representatives from these groups have been isolated from a variety of habitats such as: soils and rhizospheres [13–15], drains, fresh and salt water [16–20], plants [21, 22], and in clinical settings [23–25]. Nevertheless, despite substantial overlap in the environments from which they have been isolated, they differ in the resources they are able to use [26, 27], in their temperature tolerances [26, 28], and they have been found to differ in their relative abundances in some types of environments. For example, in a previous study in which we used a similar sampling approach to that described here, we found P. putida group strains to be more common than P. fluorescens group strains in household sites ; Igbinosa et al  found more P. putida than P. fluorescens in their freshwater and wastewater samplings; and Negi et al  isolated substantially more P. fluorescens than P. putida in their collection of cold tolerant Himalayan rhizosphere Pseudomonas strains.
In this study we used culture-based methods to characterize the distributions of members of these two species groups in a human household setting. Recovery of P. aeruginosa from this sampling effort is described elsewhere , and, consistent with other studies of human households, indicated that this species differs from other Pseudomonas in being almost a drain specialist in a household context [14, 29–32]. We explored a broad range of types of environments both inside and outside the home, in a longitudinal study including 15 households sampled up to 8 times each, over a 4.5-year period. This allowed us to characterize and compare seasonal variation in recovery in these two species groups in a range of types of environments. Even in the indoor built environment, recovery of some microbes has been found to vary seasonally, and is sometimes correlated with outdoor seasonal variability in distributions [6–9]. Inclusion of both indoor and outdoor sampling in this study allowed us to address the possibility that outdoor seasonality may be influencing indoor populations in these two groups of Pseudomonas.
Materials and Methods
We collected samples from 15 households in the Louisville, KY (USA) metropolitan area. One house in this study was included in a previous study , but the sampling results reported for that house in the previous study was not included in this analysis. Houses were each sampled between 3 and 8 times between October 2007 and March 2012, with over half (n = 8) being sampled 8 times (Fig 1). The interval between samplings ranged from 3.5 to 6.1 months, with a mean of 4.6 months. A total of 11,674 samples were taken, but as described below, 733 were eliminated from the analyses presented here resulting in a final data set of 10,941 samples. The raw data describing the characteristics of the sampling time and location, and whether or not each of the focal Pseudomonas, P. putida group and P. fluorescens group, were recovered was deposited in tab delimited text format in the Dryad Digital Repository (DRYAD), doi:10.5061/dryad.2s361.
First sampling season was Fall 2007. Frequently wet sites are shown in blue, drains in red, water in green, garbage and compost in orange, soils in purple. Human and animal sites are not shown. Fields shaded gray indicate that no sampling occurred.
Seven households included one child with cystic fibrosis (CF) and 8 had no CF patients. All households with a CF patient were recruited through the Pediatric CF Center at the University of Louisville; households without CF patients were recruited with flyers distributed on the University of Louisville Campus and by word of mouth. Because preliminary analyses of recovery of both P. putida group and P. fluorescens group recovery did not differ between these two types of houses (analyses not shown), the two types of houses were combined for the analyses presented here, and sites that are specific to homes with people with CF (upper respiratory samples from CF patients and samples from equipment related to treatment of CF) were removed from the dataset.
At each visit to a house, sampling similar to the method reported in Remold et al. . Within each household, between 74 and 168 samples (mean of 141) were collected, depending on the number of bathrooms, people, pets, etc., from 123 types of sites in and around the home (S1 Table). Subjects enrolled in the study were instructed not to clean the home the week before the sampling date. To minimize the risk of cross-contamination, no two households were sampled on the same day.
The study protocol was approved by the University of Louisville Biomedical Institutional Review Board (IRB #408.06). Written informed consent was obtained from all adult subjects and from a parent of all minors, and written assent was obtained from minors able to provide it. Sampling of animals was approved by the University of Louisville Institutional Animal Care and Use Committee (IACUC #10093).
Samples were collected with sterile swabs pre-moistened with phosphate-buffered saline. Surfaces were sampled at locations most likely to have frequent contact with human skin (i.e., knobs, buttons, etc.). Soils were sampled by inserting a swab 1–2 inches from the surface, collecting soil on the swab. All drains were swabbed within the first 1–2 inches from the top of the opening, and running water sample were taken by wetting swabs immediately after opening the tap. Some of the subjects or their parents/guardians also collected human fecal and genital samples; these were excluded from all analyses presented here due to low sample size and low recovery rates of P. putida and P. fluorescens group isolates.
All swabs were streaked onto Pseudomonas isolation agar (PIA, 10 g gelatin peptone, 10 g meat peptone, 10 g potassium sulfate, 1.4 g magnesium chloride, 25 mg Irgasan, 20 ml glycerol, 13.6 g agar, 1 L water; based on ) at the homes. PIA was used in this study because the large number of samples taken and the expected relatively low rate of recovery from some of our environments necessitated a selection that could be used to allow the growth of most Pseudomonas while eliminating most Pseudomonas-negative swabs without molecular identification. We note however that PIA has limitations with respect to both these objectives Ghyselinck et al , have found that the diversity of Pseudomonas that grow on PIA is lower than that obtained on the non-selective media TSA and PDA, and many non-Pseudomonas were obtained from our PIA plates.
Plates were transported on ice and incubated for 48 hours at 28°C immediately upon return to the laboratory. Plates without growth were incubated for an additional 24 hours before being scored as Pseudomonas negative. We chose incubation at 28°C for three reasons. First, other studies suggest that a number of Pseudomonas have wide ranges of temperature tolerance, with 25–30°C permitting growth of all studied strains [26, 35]. Second, preliminary experiments in which we isolated strains from household sites at both 28°C and 37°C yielded more strains at the lower temperature, and whereas all strains isolated at 37°C also grew at 28°C, the reverse was not true. Others have also shown that Pseudomonas isolated from cold sites (0–15°C) can be recovered at 25°C . Finally, among our non-host associated sites, all but a small minority (those associated with refrigerators and some basements and cellars) either spend most of the year at temperatures between 25 and 30°C (most indoor samples), or experience temperatures in this range regularly for at least part of the year (outdoor samples)
Where growth occurred, a single colony from each plate was picked randomly, re-streaked onto PIA. Following isolation on PIA and successful growth when re-streaked on PIA, some non-Pseudomonas were excluded using a MacConkey screen for lactose fermentation (MacConkey agar: 17 g pancreatic digest of gelatin, 1.5 g pancreatic digest of casein,1.5 g peptic digest of animal tissue,10 g lactose,1.5 g bile salts mixture, 5 g sodium chloride, 0.03 g neutral red, 0.001 g crystal violet, 13.5 g agar, 1 L water, based on ). We also employed a PCR-based screen using Pseudomonas-specific primers . Where multiple colony morphologies grew from a single sample, one of each was frozen for further analysis. From among these cases there were 22 instances in which a member of two different species groups were identified from the same sample. Of these, 12 are cases in which a P. putida group and a P. fluorescens group isolate were obtained from the same sample.
Strains identified as candidate Pseudomonas using the selections and screens described above were identified using a combination of Pseudomonas specific primers  and 16S rDNA sequencing (either on its own, or to follow up on isolates identified as Pseudomonas using genus-specific primers). For sequence-based identification, fragments amplified with universal bacterial primers (8f and 1492r) were sequenced with a 1401r primer or 8f primer . Sequences at least 500bp in length were then compared to the database Bioinfo 1200 nucleotide . Identifications were then made a second time with the database EzTaxon  to determine consistency across databases. The species identifications made with the Bioinfo 1200 nucleotide database were used only as a step in assigning each strain to a species group as defined by Anzai et al . Isolates that were identified by comparison to this database as members of species not classified by Anzai et al  were assigned to a species group by generating a provisional assignment of the unclassified species as follows: the 16S type sequence of the species to which the isolate was assigned was blasted against the NCBI database. The unclassified species was then assigned to the species group of the most similar Anzai et al. -classified species among the top blast hits. We obtained four strains with best-match species that Anzai et al  did not group, and there were two sequences of insufficient quality to assign to a group (the latter were treated as missing data with respect to P. putia and P. fluorescens group recovery). The partial 16S rDNA sequence used to identify all Pseudomonas obtained from the samples included in this study are available from GenBank under accession numbers KP452510—KP453692.
Assignment of all 251 strains to the P. fluorescens group was consistent between the two databases, Bioinfo 1200 nucleotides and EzTaxon. However, among the 668 strains classified in P. putida group there were 33, collected from a range of houses and environment types, that were all identified by Bioinfo 1200 as P. fulva, a species in the P. putida group, but by EzTaxon as members a number of different species that we had assigned to P. aeruginosa group through the method described above. In many cases, the 10 most similar sequences identified by both Bioinfo 1200 and EzTaxon included members of both P. putida and P. aeruginosa groups. We there fore performed all statistical analyses twice, with and without including these 33 strains among the P. putida group isolates. Because few qualitative differences emerged between these parallel analyses, only those conducted on the dataset including the 33 strains within P. putida group are presented here.
For analyses and for graphing we created the predictor variable “environment type” by binning the 123 types of sites sampled (S1 Table). The variable “sampling period” was created by binning sampling dates into season-year combinations, (e.g. Fall 2007), with which seasons defined as follows: winter, December-February; spring, March-May; summer, June-August; fall, September-November. A single sampling intended to occur in the fall but which had to be rescheduled to December 7 was coded as Fall for analysis to distinguish it from the sampling of the same house that occurred later that winter.
In parallel models, we fitted presence or absence of P. putida group or P. fluorescens group as the binary response variable, environment type and sampling period as fixed predictor variables, house and its interaction with environment and sampling period as random factors, and we accommodated repeated samplings of individual sampled sites at different sampling periods using an autoregressive order 1 (AR(1)) R-side covariance structure, using a general linear mixed model (PROC GLIMMIX, SAS / STAT 9.3 ). An analogous model was used to explore differences in relative recovery between the two species groups, P. putida and P. fluorescens. In this model we considered those sites from which either P. putida group or P. fluorescens group was recovered, and looked at the probability of recovering one vs. the other. For this analysis we therefore excluded all sites from which both isolates were found, those from which only members of other species groups were found, and those from which no Pseudomonas were found. We ran separate models using the data set as a whole, and using the outdoor and indoor subsets of the data set. Additional sub-models describe only spring samplings, and only fall samplings. Some random interaction terms were excluded from some sub-models when their inclusion caused models to fail to converge. Also, because categories for which no recoveries occur cause these models to fail to converge, some houses and/or sampling periods were excluded from some models.
Hypotheses regarding seasonal differences were tested using linear combinations of least squares means of sampling periods. The interaction between sampling period and environment type was not tested because there were many combinations for which no P. putida group and/or P. fluorescens group was obtained, and this situation causes the models used here to fail to converge. Random factors were tested using Wald Z tests, and least squares means estimates were used to generate estimated probabilities of recovery, odds of recovery, and odds ratios.
Overview of the strain collection
Pseudomonas species were recovered from every season, every house, and every environment type sampled (S1 Table). Out of 10,941 samples, 1,056 (9.7%) yielded Pseudomonas isolates. Twenty-two samples from which more than one morphologically distinguishable colony types were collected yielded isolates from two different species groups, resulting in a total of 1,078 clearly distinct Pseudomonas isolates in our collection. Of these 62% were from the P. putida group and 23% were from the P. fluorescens group. Recovery from the next most common species group, P. aeruginosa group, made up only 13% of isolates, and the remaining 2% of isolates (15 in total) included representatives from a number of species groups as well as some strains not assigned to any species group. Patterns of recovery of P. putida and P. fluorescens groups are highly variable across houses and sampling periods
Recovery of P. putida group or P. fluorescens group in consecutive samplings was significantly correlated in all but one model evaluated (Fig 1, AR(1) covariance terms in Tables 1–3). Nevertheless, there was also significant stochasticity in the distribution of both groups. Notably, most models detected strong sampling period-specific and/or environment type-specific among-house variation in recovery (Tables 1–4). What follows describes the additional variation attributable to the fixed predictor variables season and environment type, which exists despite this very strong variability in recovery.
Seasonal differences in recovery of P. putida and P. fluorescens groups
We used linear combinations of least squares means of sampling periods to generate estimated recovery rates by season and to test for differences among them. Overall, we found that P. putida group recovery was approximately twice as high in summer and fall than in winter and spring (Table 1, Fig 2A). In contrast, the probability of P. fluorescens group was significantly more likely to be recovered in the spring than summer, with winter and fall intermediate (Table 1, Fig 2B). Considering those samples from which one of the two was recovered, we found that odds of recovering P. putida group rather than P. fluorescens group was 5.5–6 in fall and summer. However, in winter and spring, there was no significant difference in the likelihood of recovering one species group vs. the other (Table 1, Fig 2C).
Estimated probability of recovery by season from general linear mixed model describing recovery across all sites, sampling periods and houses (See Table 1), of (A) P. putida group, (B) P. fluorescens group, and (C) odds of recovering P. putida group (as opposed to P. fluorescens group), from among sites where one or the other was recovered. Error bars indicate 95% confidence intervals. Levels that do not share a letter are significantly different at p<0.05 after Bonferroni adjustment for multiple comparisons.
We used additional models to explore the degree to which seasonal differences were driven by outdoor samplings. Outdoor recovery of both groups as well as their relative recoveries were consistent with results of analyses of the data overall. P. putida group recovery from outdoor samples was highest in summer and fall (Table 2, Fig 3A), and although there were no significant differences among seasons in outdoor recovery of P. fluorescens group the overall trend of higher recovery in spring relative to summer persisted (Table 2, Fig 3B). Among outdoor samples from which one of the two was recovered P. putida group was more likely to be recovered in summer and fall and P. fluorescens group was more likely to be recovered in winter and spring, though only difference between fall and winter is statistically significant (Table 2, Fig 3C).
Estimated probability of recovery by season from general linear mixed model describing recovery across outdoor (A-C) and indoor (D-F) sites (See Table 2), of (A, D) P. putida group, (B, E) P. fluorescens group strains, and (C, F) odds of recovering P. putida group (as opposed to P. fluorescens group), from among sites where one or the other was recovered. Levels that do not share a letter are significantly different at p<0.05 after Bonferroni adjustment for multiple comparisons. Error bars indicate 95% confidence intervals. Comparisons in panels without levels indicated yielded no significant differences after correction for multiple comparisons.
We then considered only indoor samples and samples taken from humans and pets and found that indoor recovery was also consistent with trends in the dataset as a whole. Indoor sites were significantly more likely to yield P. putida group in fall and summer than in spring, with winter intermediate (Table 3, Fig 3D), and the non-significant trend of highest P. fluorescens group recovery in spring than in other seasons remains (Table 3, Fig 3E). Trends in relative recovery were also consistent the dataset as a whole in that the dominance of P. putida group is estimated to be strongest in summer and fall (Table 3, Fig 3F).
Differences among types of environments in recovery of P. putida and P. fluorescens groups
In the data set as a whole, P. putida group recovery from soils was almost four times higher than recovery from drains, which in turn was also higher than recovery from most other environment types (Table 1, Fig 4). In contrast, P. fluorescens group recovery was more similar across environment types, though as was the case for P. putida group, soil recovery was higher than recovery from other environment types (Table 1, Fig 4). Although recovery from human and pet upper respiratory and skin sites was low for both species groups, some P. fluorescens group and P. putida group isolates were recovered from all of these four environment types. Considering only samples from which one of the two was recovered, P. putida group was approximately 8 to 10 times more likely in soil samples than P. fluorescens group (Table 1, Fig 4), but in a number of other environment types estimated recovery of P. putida and P. fluorescens groups relative to the other was not statistically different from equal.
Estimated probability of recovery by environment type from general linear mixed model describing recovery across all sites (See Table 1), of (A) P. putida group, (B) P. fluorescens group strains, and (C) odds of recovering P. putida group (as opposed to P. fluorescens group), from among sites where one or the other was recovered. Error bars indicate 95% confidence intervals. Levels that do not share a letter are significantly different at p = 0.05 after Bonferroni adjustment for multiple comparisons.
The ranking of environments with respect to probability of yielding of P. putida group was consistent indoors vs. outdoors, though overall recovery for both species groups was higher outdoors (Tables 2 and 3, Fig 5A and 5D). For P. fluorescens group there were few (indoors, Table 3, Fig 5E), or no (outdoors, Table 2, Fig 5B) significant differences in recovery among environment types. Among outdoor samples there were no significant differences in the odds of recovering one species group vs. the other (Table 2, Fig 5C), but indoors the pattern of variability in relative rates of recovery across environments was highly consistent with the sample overall (Table 3, Fig 5F).
Estimated probability of recovery by environment type from general linear mixed model describing recovery across outdoor (A-C, See Table 2) and indoor (D-F, See Table 3) sites, of (A,D) P. putida group, (B,E) P. fluorescens group strains, and (C,F) odds of recovering P. putida group (as opposed to P. fluorescens group), from among sites where one or the other was recovered. Error bars indicate 95% confidence intervals. Levels that do not share a letter are significantly different at p = 0.05 after Bonferroni adjustment for multiple comparisons. Comparisons in panels without levels indicated yielded no significant differences after correction for multiple comparisons.
It was not possible to test for season-specific differences in recovery among environment types in our full model due to an excess of combinations of these factors from which no isolates were obtained. We therefore ran separate sub-models on samplings occurring in fall and in spring only. These models excluded human and animal-associated samples due to low recovery rates. For P. putida group, the relative ranking of the environment types was consistent between fall and spring, though estimated recoveries in the spring were on average less than half those of the fall (Fig 6A vs. 6D), and the reverse was true for P. fluorescens group (Fig 6B vs. 6E). Whereas in the fall, the estimated likelihood of recovering P. putida group was significantly greater than that of recovering P. fluorescens groups (with estimates ranging from 4.5 times more likely on surfaces, to over 36 times more likely in soils), in the spring recovery from all habitat types except soil and drains was not different from equal, with estimates suggesting P. fluorescens may be more common than P. putida in many of these types of sites (Table 4, Fig 6).
Estimated probability of recovery by environment type from general linear mixed model describing recovery across sites sampled in spring (A-C) and fall (D-F), of (A,D) P. putida group, (B,E) P. fluorescens group strains, and (C,F) odds of recovering P. putida group (as opposed to P. fluorescens group), from among sites where one or the other was recovered. Error bars indicate 95% confidence intervals. Levels that do not share a letter are significantly different at p = 0.05 after Bonferroni adjustment for multiple comparisons. Comparisons in panels without levels indicated yielded no significant differences after correction for multiple comparisons.
In this study we took 10,941 samples taken from over 100 types of sites in 15 homes, across 8 years, with the goal of investigating the temporal and spatial variation in distribution patterns of P. putida and P. fluorescens group Pseudomonas. We found striking differences in rates of recovery between P. putida and P. fluorescens groups overall, as well as differences in their patterns of seasonal variation, and differential use of environment types. While it is unlikely that artifacts of our culture-based our sampling methods could generate the temporal and spatial patterns observed, two factors result in our absolute recovery rates being biased toward underestimation of the absolute rate of Pseudomonas occurrence. First, our culture media, PIA, excludes some Pseudomonas , and this media and/or our culture temperature may cause underestimation and/or bias in detection of P. putida and P. fluorescens groups through exclusion of genotypes. Second, because we identified only one strain per colony morphology on each plate, growth of non-Pseudomonas on our PIA plates likely also contributed to underestimation of Pseudomonas occurrence.
With this limitation in mind, we note that our finding that Pseudomonas were obtained from all environment types sampled, and yet were below the limit of detection in most individual samples suggests they can occur in most sites, but are either rare or undergo frequent extinction and recolonization. This hypothesis is consistent with the results of Dunn et al (2013), who found that the mean relative abundance of all Pseudomonales among high throughput sequence reads obtained from nine household site types was low (5.5–10.8% in kitchen sites, and 1.2–3.3% in other types of sites). Taken together these studies suggest that any one species or species group of Pseudomonas may be sufficiently rare that high throughput sequence approach using Pseudomonas-specific primers for a region that yields greater taxonomic resolution, such as rpoD,  may be necessary to accurately quantify abundance at the species level. This approach will also aid in determining whether populations are small but permanent or go through rounds of extinction and recolonization. Resolving these issues will contribute to uncovering the relative roles of possible drivers of differential recovery discussed below.
The overall difference in relative recovery of P. putida and P. fluorescens groups is consistent with previous results . It occurred despite the fact that members of both species groups can use a broad range of resources, and the fact that strains of both have been isolated from many different environments, including from all of the environment types sampled in this study . Therefore, although they both use many types of household sites, in a given sampled area, P. putida group members are either more abundant or easier to obtain than P. fluorescens group members. If the former is important, it may reflect difference in the density to which these organisms commonly grow, perhaps as a result in differences in their interactions with the substrate on which they are growing, or with other members of the community. As discussed below, because these two species group differ with respect to which season they are most abundant, studies that sample in just one season may also underestimate P. fluorescens group use of household habitats.
Our findings also indicate that household sites differ in their suitability for Pseudomonas, and that they do so in a taxon-specific way. Members of both species groups were recovered from all environment types, and both were more frequently recovered from soils than from other sites, but whereas soil recovery, followed by drain recovery substantially exceeded recovery in other environment types for P. putida group, for P. fluorescens group recovery was much more similar among environment types. The relative rate of recovery of P. putida group vs. P. fluorescens group also varied among environment types, with soils being significantly more biased toward P. putida group than transiently wet sites and surfaces (Tables 1–4, Figs 4–6). P. aeruginosa, a member of the P. aeruginosa species group exhibits yet another distinct recovery pattern: in a household setting, it is most commonly found in drains [14, 29–32]. These findings are consistent with that of Remold et al , who found that some Pseudomonas species were over- or under- represented in some environments relative to their representation across a range of household environments.
There were distinct seasonal trends in the distributions of these two species, with P. putida group recovery highest in fall and summer, and P. fluorescens group recovery highest in spring and to a lesser extent in winter, resulting in much greater dominance of P. putida group over P. fluorescens group in fall than in spring (Tables 1–4). Higher P. fluorescens group recovery in spring and winter is consistent with this species being psychrophilic, as a number of strains of this species have been found to be [15, 26, 42, 43]. While the driver of the substantial decrease in P. putida group recovery in spring relative to fall is unclear, it is consistent with results from South Carolina stream samples, from which more P. putida recovery was also higher in the fall than in winter or spring .
For both species groups, seasonal patterns outdoors and indoors were similar. The finding that sampling inside homes showed seasonal variation (Table 3, Figs 3 and 4) is consistent with the observation of seasonal variation in a number of studies of the built environment [6–9]. Nearby external inputs have been shown to play an important role in shaping the microbial communities found in the built environment, including in homes (e.g. [7, 44]). While it is possible that in the homes we studied the similarity in recovery trends indoors and outdoors is driven by frequent movement of Pseudomonas from outside the home to inside, abiotic factors inside the home, or interactions with other organisms that exhibit seasonal variation of movement from outdoors to indoors could also play a role.
In conclusion, we found striking differences between the P. putida and P. fluorescens groups in their recovery overall, and across seasons and environment types. This could be driven by a number of factors . First, differences in dispersal opportunities could affect access to household sites. This might, for example, explain higher overall recovery outdoors than indoors. Second, species sorting driven by differences in how members of the two species groups interact with features of the abiotic environment may influence relative recoveries. For example, there may be differences between P. putida group and P. fluorescens group in sensitivity to stresses such as extreme temperatures or desiccation. If so, this could generate seasonal differences or differences in relative abundance among sites. In addition, direct competition could drive niche partitioning between members of these species groups. Finally, interactions with other community members, such as phages (both species groups as well as P. aeruginosa have been shown to be vulnerable to phage [46–48]) or predators such as fungi or amoebae could contribute to the differential use of household environments by these two species groups. Regardless of the underlying mechanism, our study shows that putative generalists with apparently broad habitat use patterns may nevertheless exhibit complex differences in habitat use, and highlights the importance of considering seasonal variation even in studies of indoor environments.
S1 Table. Sites sampled in the homes by environment type bin.
Upper respiratory sites were sampled in people with and without cystic fibrosis (CF); only those from people without CF were included in the analyses presented here. CF equipment sites were obtained from a subset of houses and were also not included in analyses presented here. Fecal and genital samples were obtained from a subset of volunteers and were not included in analyses presented here.
We thank our volunteer families for allowing us into their homes and lab members who participated in sampling and identifications. We thank Nemr Eid for assistance in recruiting CF families, and Nemr Eid, Margaret Carreiro, Sarah Emery, Jim Graham, Mike Perlin Eileen Remold-O’Donnell and four anonymous reviewers for helpful comments.
Conceived and designed the experiments: SKR. Performed the experiments: SKR MEP-G MTF TCH. Analyzed the data: SKR MEP-G. Wrote the paper: SKR MEP-G.
- 1. Smit E, Leeflang P, Gommans S, van den Broek J, van Mil S, Wernars K. Diversity and seasonal fluctuations of the dominant members of the bacterial soil community in a wheat field as determined by cultivation and molecular methods. Appl Environ Microbiol. 2001;67(5):2284–91. pmid:ISI:000168488400042.
- 2. Lipson DA, Schmidt SK. Seasonal changes in an alpine soil bacterial community in the Colorado Rocky Mountains. Appl Environ Microbiol. 2004;70(5):2867–79. pmid:ISI:000221340400041.
- 3. Smalla K, Wieland G, Buchner A, Zock A, Parzy J, Kaiser S, et al. Bulk and rhizosphere soil bacterial communities studied by denaturing gradient gel electrophoresis: plant-dependent enrichment and seasonal shifts revealed. Appl Environ Microbiol. 2001;67(10):4742–51. pmid:11571180
- 4. Jasser I, Królicka A, Jakubiec K, Chróst RJ. Seasonal and spatial diversity of Picocyanobacteria community in the Great Mazurian Lakes derived from DGGE analyses of ITS region of rDNA and cpcBA-IGS markers. J Microbiol Biotechn. 2013;23(6):739–49. pmid:ISI:000321599500001.
- 5. Giovannoni SJ, Vergin KL. Seasonality in ocean microbial communities. Science. 2012;335(6069):671–6. pmid:ISI:000300047100041.
- 6. Moschandreas DJ, Pagilla KR, Storino LV. Time and space uniformity of indoor bacteria concentrations in Chicago area residences. Aerosol Sci Tech. 2003;37(11):899–906. pmid:ISI:000184457100004.
- 7. Rintala H, Pitkaeränta M, Toivola M, Paulin L, Nevalainen A. Diversity and seasonal dynamics of bacterial community in indoor environment. BMC Microbiol. 2008;8:56. pmid:ISI:000255650600001.
- 8. Adams RI, Miletto M, Lindow SE, Taylor JW, Bruns TD. Airborne bacterial communities in residences: similarities and differences with fungi. PloS One. 2014;9(3):e91283. pmid:24603548
- 9. Frankel M, Bekö G, Timm M, Gustavsen S, Hansen EW, Madsen AM. Seasonal variations of indoor microbial exposures and their relation to temperature, relative humidity, and air exchange rate. Appl Environ Microbiol. 2012;78(23):8289–97. pmid:23001651
- 10. Southwood T. Habitat, the templet for ecological strategies? The Journal of Animal Ecology. 1977:337–65.
- 11. Roughgarden J. Evolution of niche width. The American Naturalist. 1972;106(952):683–718.
- 12. Anzai Y, Kim H, Park JY, Wakabayashi H, Oyaizu H. Phylogenetic affiliation of the pseudomonads based on 16S rRNA sequence. Int J Syst Evol Microbiol. 2000;50:1563–89. pmid:ISI:000088405900021.
- 13. Cho JC, Tiedje JM. Biogeography and degree of endemicity of fluorescent Pseudomonas strains in soil. Appl Environ Microbiol. 2000;66(12):5448–56. pmid:ISI:000167112400052.
- 14. Remold S, Brown C, Farris J, Hundley T, Perpich J, Purdy M. Differential habitat use and niche partitioning by Pseudomonas species in human homes. Microb Ecol. 2011;62(3):505–17. pmid:21503776
- 15. Negi YK, Prabha D, Garg SK, Kumar J. Genetic diversity among cold-tolerant fluorescent Pseudomonas isolates from Indian Himalayas and their characterization for biocontrol and plant growth-promoting activities. J Plant Growth Regul. 2011;30(2):128–43.
- 16. Alouache S, Kada M, Messai Y, Estepa V, Torres C, Bakour R. Antibiotic resistance and extended-spectrum β-Lactamases in isolated bacteria from seawater of Algiers beaches (Algeria). Microbes Environ. 2012;27(1):80–6. pmid:22095134
- 17. Vullo DL, Ceretti HM, Hughes EA, Ramírez S, Zalts A. Indigenous heavy metal multiresistant microbiota of Las Catonas stream. Environ Monit Assess. 2005;105(1–3):81–97. pmid:15952533
- 18. Lemke M, Leff L. Bacterial populations in an anthropogenically disturbed stream: comparison of different seasons. Microb Ecol. 1999;38(3):234–43. pmid:10541785
- 19. Marinho PR, Moreira APB, Pellegrino FL, Muricy G, Bastos M do C, Santos KR, et al. Marine Pseudomonas putida: a potential source of antimicrobial substances against antibiotic-resistant bacteria. Mem Inst Oswaldo Cruz. 2009;104(5):678–82. pmid:19820824
- 20. Igbinosa IH, Nwodo UU, Sosa A, Tom M, Okoh AI. Commensal Pseudomonas species isolated from wastewater and freshwater milieus in the Eastern Cape Province, South Africa, as reservoir of antibiotic resistant determinants. Int J Env Res Public Health. 2012;9(7):2537–49.
- 21. Wu Z, Kan F, She Y-M, Walker V. Biofilm, ice recrystallization inhibition and freeze-thaw protection in an epiphyte community. Appl Biochem Microbiol. 2012;48(4):363–70.
- 22. Dimartino M, Panebianco S, Vitale A, Castello I, Leonardi C, Cirvilleri G, et al. Occurrence and pathogenicity of Pseudomonas fluorescens and P. putida on tomato plants in Italy. J Plant Pathol. 2011:79–87.
- 23. Molina L, Udaondo Z, Duque E, Fernández M, Molina-Santiago C, Roca A, et al. Antibiotic resistance determinants in a Pseudomonas putida strain isolated from a hospital. PloS One. 2014;9(1):e81604. pmid:24465371
- 24. Jacquier H, Carbonnelle E, Corvec S, Illiaquer M, Le Monnier A, Bille E, et al. Revisited distribution of nonfermenting Gram-negative bacilli clinical isolates. Eur J Clin Microbiol Infect Dis. 2011;30(12):1579–86. pmid:21509476
- 25. Wong V, Levi K, Baddal B, Turton J, Boswell TC. Spread of Pseudomonas fluorescens due to contaminated drinking water in a bone marrow transplant unit. J Clin Microbiol. 2011;49(6):2093–6. pmid:21450958
- 26. Stanier R, Palleroni N, Doudoroff M. The aerobic pseudomonads a taxonomic study. Journal of General Microbiology. 1966;43(2):159–271. pmid:5963505
- 27. Clarke PH. The metabolic versatility of pseudomonads. Antonie Van Leeuwenhoek. 1982;48(2):105–30. pmid:6808915
- 28. Blazevic DJ, Koepcke MH, Matsen JM. Incidence and identification of Pseudomonas fluorescens and Pseudomonas putida in the clinical laboratory. Applied microbiology. 1973;25(1):107–10. pmid:4631431
- 29. Purdy-Gibson ME, France M, Hundley TC, Eid N, Remold SK. Pseudomonas aeruginosa in CF and non-CF homes is found predominantly in drains. Journal of Cystic Fibrosis. 2015;14(2015):341–346. http://dx.doi.org/10.1016/j.jcf.2014.10.008.
- 30. Ojima M, Toshima Y, Koya E, Ara K, Kawai S, Ueda N. Bacterial contamination of Japanese households and related concern about sanitation. Int J Environ Health Res. 2002;12(1):41–52. pmid:ISI:000173645000004.
- 31. Regnath T, Kreutzberger M, Illing S, Oehme R, Liesenfeld O. Prevalence of Pseudomonas aeruginosa in households of patients with cystic fibrosis. Int J Hyg Environ Health. 2004;207(6):585–8. pmid:ISI:000227234400012.
- 32. Schelstraete P, Van Daele S, De Boeck K, Proesmans M, Lebecque P, Leclercq-Foucart J, et al. Pseudomonas aeruginosa in the home environment of newly infected cystic fibrosis patients. Eur Respir J. 2008;31(4):822–9. pmid:ISI:000254866300020.
- 33. Atlas RM. Handbook of microbiological media: CRC press; 2010.
- 34. Ghyselinck J, Coorevits A, Van Landschoot A, Samyn E, Heylen K, De Vos P. An rpoD gene sequence based evaluation of cultured Pseudomonas diversity on different growth media. Microbiology. 2013;159(Pt 10):2097–108. pmid:23920133
- 35. Tsuji A, Kaneko Y, Takahashi K, Ogawa M, Goto S. The effects of temperature and pH on the growth of eight enteric and nine glucose non-fermenting species of gram-negative rods. Microbiol Immunol. 1982;26(1):15–24. pmid:7087800
- 36. Doulgeraki AI, Nychas G-JE. Monitoring the succession of the biota grown on a selective medium for pseudomonads during storage of minced beef with molecular-based methods. Food Microbiol. 2013;34(1):62–9. pmid:23498179
- 37. Spilker T, Coenye T, Vandamme P, LiPuma JJ. PCR-Based assay for differentiation of Pseudomonas aeruginosa from other Pseudomonas species recovered from cystic fibrosis patients. J Clin Microbiol. 2004;42(5):2074–9. pmid:ISI:000221424100032.
- 38. Weisburg WG, Barns SM, Pelletier DA, Lane DJ. 16S ribosomal DNA amplification for phylogenetic study. J Bacteriol. 1991;173(2):697–703. pmid:ISI:A1991ET44600034.
- 39. Croce O, Chevenet F, Christen R. A new web server for the rapid identification of microorganisms. Journal of Microbial & Biochemical Technology. 2010;2(3):084–8.
- 40. Chun J, Lee J- H, Jung Y, Kim M, Kim S, Kim BK, et al. EzTaxon: a web-based tool for the identification of prokaryotes based on 16S ribosomal RNA gene sequences. Int J Syst Evol Microbiol. 2007;57(10):2259–61. pmid:17911292
- 41. SAS Institute I. SAS / STAT 9.3 User's Guide. Cary, NC.: SAS Institute, Inc.; 2011.
- 42. Ingraham J. Growth of psychrophilic bacteria. J Bacteriol. 1958;76(1):75. pmid:13563393
- 43. Guillou C, Guespin-Michel J. Evidence for two domains of growth temperature for the psychrotrophic bacterium Pseudomonas fluorescens MF0. Appl Environ Microbiol. 1996;62(9):3319–24. pmid:8795221
- 44. Dunn RR, Fierer N, Henley JB, Leff JW, Menninger HL. Home life: factors structuring the bacterial diversity found within and between homes. Plos One. 2013;8(5). pmid:ISI:000320362700107.
- 45. Ricklefs RE, Jenkins DG. Biogeography and ecology: towards the integration of two disciplines. Philos T R Soc B. 2011;366(1576):2438–48. pmid:ISI:000292819200010.
- 46. Sillankorva S, Neubauer P, Azeredo J. Pseudomonas fluorescens biofilms subjected to phage phiIBB-PF7A. BMC Biotechnol. 2008;8(1):79.
- 47. Shaburova OV, Krylov SV, Veiko VP, Pleteneva EA, Burkal'tseva MV, Miroshnikov KA, et al. Search for destruction factors of bacterial biofilms: Comparison of phage properties in a group of Pseudomonas putida bacteriophages and specificity of their halo-formation products. Russ J Genet. 2009;45(2):161–70. pmid:ISI:000263781700005.
- 48. Knezevic P, Obreht D, Curcin S, Petrusic M, Aleksic V, Kostanjsek R, et al. Phages of Pseudomonas aeruginosa: response to environmental factors and in vitro ability to inhibit bacterial growth and biofilm formation. J Appl Microbiol. 2011;111(1):245–54. pmid:ISI:000292884000025.