Skip to main content
NCBI home page
The ISME Journal logoLink to The ISME Journal
. 2017 Jun 23;11(11):2439–2451. doi: 10.1038/ismej.2017.99

Linking the development and functioning of a carnivorous pitcher plant’s microbial digestive community

David W Armitage 1,2,*
PMCID: PMC5649175  PMID: 28644442

Abstract

Ecosystem development theory predicts that successional turnover in community composition can influence ecosystem functioning. However, tests of this theory in natural systems are made difficult by a lack of replicable and tractable model systems. Using the microbial digestive associates of a carnivorous pitcher plant, I tested hypotheses linking host age-driven microbial community development to host functioning. Monitoring the yearlong development of independent microbial digestive communities in two pitcher plant populations revealed a number of trends in community succession matching theoretical predictions. These included mid-successional peaks in bacterial diversity and metabolic substrate use, predictable and parallel successional trajectories among microbial communities, and convergence giving way to divergence in community composition and carbon substrate use. Bacterial composition, biomass, and diversity positively influenced the rate of prey decomposition, which was in turn positively associated with a host leaf’s nitrogen uptake efficiency. Overall digestive performance was greatest during late summer. These results highlight links between community succession and ecosystem functioning and extend succession theory to host-associated microbial communities.

Introduction

Although the capacity for community composition to mediate ecosystem processes is widely recognized (Hooper et al., 2005), few theoretical (Finn, 1982; DeAngelis, 1992; Loreau, 1998) and empirical studies (Fisher et al., 1982; Schmidt et al., 2007) have investigated community-ecosystem linkages along natural successional gradients. Ecosystem development theory (Odum, 1969) seeks to explain temporal variation in ecosystem properties in terms of community successional turnover. Central to this theory is the prediction that successional turnover can influence elemental cycling rates leading to a coupling of community composition and ecosystem processes through time (Odum, 1969; Huston and Smith, 1987; DeAngelis, 1992; Loreau, 1998). It is worth noting that modern succession theory does not assume directionality toward a stable equilibrium (or ‘climax’), but instead recognizes that the temporal trajectories of ecosystems can vary due to the relative influences of general ecological processes (Meiners et al., 2015). Although these predictions have not been immune to critique on both theoretical and empirical grounds, adequately replicated tests in natural communities remain scarce.

The natural microcosms of host-associated microbial communities offer a number of unique advantages for testing ecosystem development hypotheses. First, microbiota can enable identifiable and measureable functions for their hosts (Bäckhed et al., 2005; Lugtenberg and Kamilova, 2009). Next, the habitats being colonized are often nearly identical among closely related individuals, permitting repeated, independent observations of ecosystem development. Finally, the successional dynamics of host-associated microbiota frequently operate over time scales proportional to the host’s lifespan, which can manifest as large shifts in community composition and function over relatively short time periods.

This study uses the microbial digestive communities in developing leaves of the pitcher plant Darlingtonia californica (Sarraceniaceae) (Figure 1a) to test the following hypotheses linking community succession to ecosystem function (Figures 1b–e): First, alpha diversity will either asymptotically increase or be unimodal over the host leaf’s lifespan as taxa are recruited from the regional pool and subsequently persist or are excluded by superior competitors (Odum, 1969; Loucks, 1970; Auclair and Goff, 1971; Connell and Slatyer, 1977; Fierer et al., 2010). Consequently, trait diversity (for example, biochemical pathways, C-substrate use) is also expected to increase as succession proceeds (Odum, 1969). Second, rates of biomass production should decrease over time, as growth-limiting nutrients are lost from the system and/or stored in living biomass—this should manifest as a logistic-like biomass curve (Odum, 1969; Vitousek and Reiners, 1975; Fierer et al., 2010). Third, beta diversity will increase over time if environmental differences among pitchers cause spatially-variable selection or drift, or decrease over time if different leaves constitute similar selective environments (Christensen and Peet, 1984; Dini-Andreote et al., 2015). Fourth, host ecosystem properties (for example, nutrient cycling, decomposition) should increase monotonically or be unimodal, concomitant with changes in alpha diversity and biomass, as the accumulation of individuals of different species accelerates the degradation of organic material (Cardinale et al., 2007; Weis et al., 2007; Armitage, 2016). This leads to the prediction that biodiversity and biomass dynamics will set ecosystem process rates (e.g., decomposition), which, in turn, will set rates on host functioning (for example, nutrient uptake rates; Hooper et al., 2005).

Figure 1.

Figure 1

Open in a new tab

Predictions for successional patterns in Darlingtonia leaves. (a) Conceptual model for the interactions between host leaf (shaded oval) and its digestive food web (boxes). Dashed gray arrows denote ecological processes hypothesized to influence food web dynamics and host functioning. (b) α-diversity and living biomass are predicted to increase in pitcher leaves after opening, and eventually either saturate or decrease, consistent with observations made across a variety of ecosystems (Odum, 1969; Loucks, 1970; Auclair and Goff, 1971; Vitousek and Reiners, 1975; Connell and Slatyer, 1977; Peet and Christensen, 1988; Alday et al., 2011). (c) Compositional differences among leaf communities (β-diversity) may either decrease or increase depending on whether selection is homogenous or variable among leaves (Christensen and Peet, 1984; Dini-Andreote et al., 2015; Meiners et al., 2015). (d) Ecosystem or host function is anticipated to be unimodal or saturating over a successional gradient (van Ruijven and Berendse, 2005; Cardinale et al., 2007; Weis et al., 2007; Armitage, 2016), a pattern predicted to be influenced by (e) the positive effects of α-diversity and living biomass on ecosystem function (Bell et al., 2005; Hooper et al., 2005). Dashed lines denote alternative hypotheses.

To test these hypotheses, I followed cohorts of pitcher leaves over three years and quantified their associated digestive communities through time. In addition, I measured these communities’ rates of decomposition, respiration, and their host leaves’ nitrogen uptake efficiencies. These data were used to test whether host-associated digestive communities follow general, predictable successional patterns and whether their turnover can influence a host’s ability to digest prey and sequester nutrients.

Materials and Methods

Complete documentation of the study system, data collection, and statistical analyses are provided in the Supplementary Materials and Methods.

In situ isotopic labeling of pitcher leaves

A stable isotope pulse-chase experiment was used to measure rates of decomposition and nitrogen cycling by the pitchers’ aquatic food webs. In early June 2013, I identified and tagged 50 unopened Darlingtonia pitcher leaves of equivalent age on different plants growing in a large population in the Plumas National Forest (Plumas Co., Quincy, CA, USA). Pitcher leaves remain sterile until completing their development and commencing prey capture, and each leaf has a lifespan of ~1.5 years. On the day the pitcher leaves first opened in mid-June, I fed gel capsules containing 20 sterile, 15N-enriched fruit flies (Drosophila melanogaster) to five random leaves, which were then left undisturbed for 11 days. I returned to the site to remove these 15N-labeled pitcher leaves and to feed isotope-labeled flies to 5 additional leaves belonging to the same cohort. This process was repeated every 11 days up to day 88 (mid-September), and again on day 365 (June 2014) with 10 leaves. Because the weight of enriched flies (4.25 mg) was much smaller than the average (177 mg) and standard deviations (176 mg) of natural prey masses within a leaf age class, this prey addition was unlikely to significantly overwhelm the natural variation in nutrient levels experienced by pitcher food webs. The 11-day timeframe was chosen based on preliminary data demonstrating peak N incorporation rates by lab-reared plants between 4 and 11 days after prey capture. I repeated this experiment in 2014–2015 in a nearby population of D. californica and included an additional 166-day sample. The sampled leaves were placed on ice and quickly returned to the lab.

Quantification of pitcher leaf communities through time

From each freshly-collected leaf I removed 700 μl of fluid for DNA extraction using the PowerSoil microbial DNA isolation kit (MoBio Laboratories, Inc., Carlsbad, CA, USA) and stored the extractions at −80 °C. Next, I dissected the pitcher leaves and categorized the state of fruit fly decomposition on an ordinal scale from 0 (no decomposition; flies undamaged) to 5 (completely decomposed; head capsules and wings only). I identified and enumerated all protists and living arthropods (primarily Sarraceniopus mites and Metriocnemus midge larvae) in each leaf’s fluid and interior surface under a light microscope and used epifluorescence microscopy to enumerate SYBR-Gold (Thermo Fisher Scientific, Inc., Waltham, MA, USA) stained bacterial cells and virus-like-particles bound to 0.02 μm filters. All prey detritus in a leaf was oven-dried at 60 °C and weighed.

Bacterial community sequencing

Extracted DNA was sent for PCR amplification of the 16S SSU-rRNA genes (primer set 515f/806r) and multiplexed 2 × 150 bp paired-end sequencing on the Illumina MiSeq at the Argonne National Lab Core Sequencing Facility (Lemont, IL). Sequences were deposited on the MG-RAST public server (http://metagenomics.anl.gov/) under project ID mgp14344. The QIIME bioinformatics pipeline was used to assemble and cluster reads into 97% operational taxonomic units (OTUs; Caporaso et al., 2010). I calculated each community’s alpha diversity (Shannon’s H, richness, phylogenetic) and beta diversity (Jensen-Shannon distance and weighted/unweighted UniFrac—a measure of community phylogenetic dissimilarity) using the vegan and PhyloSeq R packages, and used library size factor (LSF) normalization for all beta diversity metrics (McMurdie and Holmes, 2013; Love et al., 2014; Oksanen et al., 2015; R Development Core Team, 2015). Beta diversities for each sampling period were estimated using average inter-sample distances, and the results were unchanged when distances-to-centroid were used. I tested whether community composition changed with pitcher age using permutational analysis of variance (Anderson, 2001) on samples’ Jensen-Shannon distances (JSD) and UniFrac distances and visualized these results using PCoA plots. Results were unchanged when rarefaction was used for normalization.

To assess the generality of successional turnover in pitcher communities, I modeled OTU counts using a negative binomial generalized linear model (GLM) (Love et al., 2014). Models were fit using empirical Bayes and OTUs experiencing significant log2-fold change among time points were identified using Wald P-values. I defined the ‘successional microbiome’ as the subset of OTUs experiencing a statistically significant (α=0.01) ⩾8-fold change in abundance between any two pitcher age classes and used these OTUs to construct an abundance-weighted heat map. The predictive accuracy of this subset of OTUs was assessed by training a random forest machine learning algorithm on OTU counts from the 2013 study population and using it to predict the age of samples from the independent 2014 study population. Model accuracy was evaluated using the coefficient of determination (R2) for predicted vs observed ages along a 1:1 line. The entire bioinformatic/analytical pipeline is illustrated in Supplementary Figure 1.

Estimating microbial community traits

The Biolog GN2 microplate assay (Biolog Inc., Hayward, CA, USA) was used to measure the carbon substrate use patterns of the microbial communities from an independent collection of 11, 55 and 365-day-old pitchers (10 from each age in 2014). These time points were chosen to represent early, middle, and late-stage communities. Plates were inoculated in triplicate using 20 μm-filtered, dilute, starved communities, and incubated for 3 days at 25 °C. I regressed substrate counts against leaf age using a negative binomial GLM to determine whether the number of metabolized substrates differed by leaf community age. To visualize differences in substrate profiles between age classes, I plotted samples onto principal coordinate (PCoA) axes based on their Jaccard distances.

I used ancestral genome reconstruction implemented by the PICRUSt software (Langille et al., 2013) to predict the rRNA copy number and functional gene contents for the subset of OTUs in my samples present in the greengenes database (nearest sequenced taxon index=0.071±0.01 s.e.m.). I estimated the mean weighted rRNA copy number of each pitcher sample (Nemergut et al., 2015) and then evaluated their temporal turnover using ANOVA. Pitcher samples were then ordinated based on their predicted level 3 KEGG pathway relative abundances (Kanehisa et al., 2016) using principal components analysis (PCA) and then hierarchically clustered. I filtered KEGG pathways using ANOVA P-values (P⩽0.01) and effect sizes (η2⩾0.26) in order to identify genes and pathways (focusing primarily on enzymes involved in protein degradation and nitrogen transformation) that were predicted to be differentially enriched across time points. The predictive nature of these data precluded statistical hypothesis testing, and are treated as speculative hypotheses.

Quantification of pitcher ecosystem properties through time

Empty pitcher leaves were thoroughly rinsed, dried at 60 °C, homogenized in a bead-beater, weighed, and analyzed for 15N using an isotope ratio mass spectrometer at the UC Davis Stable Isotope Facility (Davis, CA, USA). I used the fly and leaf 15N measurements to estimate the total amount of fly-derived 15N found in a leaf’s tissue after 11 days, which is interpreted to be the host leaf’s nitrogen uptake efficiency.

To estimate each pitcher microbial community’s potential C-respiration rate, I inoculated starved, washed pellets of pitcher bacteria into deep-well plates containing 800 μl sterile medium comprised of M9 salt solution and ground cricket powder. I used the MicroResp respirometry system to measure the rates of CO2-C respired from cultures over three days at 25 °C. These rates of CO2 respiration reflect the potential respiration rates of each pitcher’s bacterial community in a common environment.

I assessed temporal variation in pitcher ecosystem properties using ANOVA for N-uptake efficiency/carbon respiration and a multinomial logit model for the fly decomposition category (Agresti, 2013). Covariates in these models included bacterial biomass and diversity, midge larvae abundance, leaf dry weight, and leaf age. Best-fit models were identified pluralistically using a combination of R2 and small-sample adjusted Akaike Information Criterion (AICc) statistics (Burnham and Anderson, 2003). To investigate whether bacterial community composition influenced host functioning, I ran a Mantel test to assess whether pairwise Euclidean distances among samples’ N-uptake efficiencies covaried with their pairwise JSD or UniFrac dissimilarity metrics.

Verifying the effects of community structure on host function

Pitcher leaves of differing ages might physiologically regulate nitrogen uptake independent of their associated food webs, which can obscure food web effects. To account for this, I ran a field experiment to separate the effects of the food web and host leaf age on rates of N-uptake. During late July 2014 I identified 15 pitcher leaves aged 11 days, 55 days and >365 days (5 leaves of each age), intended to represent young, middle-aged, and senescing pitchers based on developmental trends observed the previous year. The fluid from these leaves was removed and mixed in equal parts to form a homogenate. 5 ml aliquots of these homogenized communities were then returned to the host plants. Additionally, 20 15N-enriched fruit flies were delivered into each leaf. I returned after 11 days to harvest and process these pitchers for N-uptake efficiency as previously described. I used ANOVA to test whether the N-uptake efficiencies of these pitchers with homogenized food webs recapitulated the N-uptake patterns from natural pitcher food webs of equivalent age from the same population.

Results

Temporal changes in the Darlingtonia food web

The dynamics of dead and living biomass were qualitatively similar to the predictions in Figure 1a. Pitcher leaves’ prey biomass varied widely among leaves of the same age, and mean prey masses quickly increased after opening and remained relatively stable throughout the plant’s lifespan (Figure 2a). Bacterial biomass also rapidly accumulated in young pitcher leaves and increased over time during the first growing season to a maximum of 1 × 1011 cells ml−1 before declining during the second growing season (Figure 2a). Virus-like particles, Sarraceniopus darlingtonae mites, and Polytomella agilis flagellates also increased in abundance during the first growing season (Figure 2a; Supplementary Figure 1). In addition to P. agilis, I detected numerous other eukaryotes, including Bodo, Monas, Petalomonas, Rhynchobodo, Chilomonas, Colpoda, Philodina and Chlamydomonas, but these taxa were observed in 10 or fewer pitcher leaves with no apparent temporal trends in occupancy or richness (Supplementary Figure 2). Likewise, I did not detect a temporal trend in bacterivore beta diversity among time points until they diverged in year 2 (Supplementary Figure 2).

Figure 2.

Figure 2

Open in a new tab

Trends in community composition during succession. (a) Insect prey biomass rapidly increased in leaves after opening and remained relatively steady throughout the remainder of the leaf’s lifespan, while bacterial and midge larval abundances steadily increased throughout leaves’ first growing season, and then sharply declined after the first year. (b) Bacterial alpha diversities increased and then leveled off in middle-aged pitcher communities, dropping slightly during year 2. (c) Conversely, leaf bacterial beta diversities decreased during the first growing season and increased at the beginning of year 2. In each graph, shared letters above groups indicate no significant pairwise differences (P>0.05). Points denote mean values±s.e.m.

Composition and convergence of pitcher bacterial communities

After quality filtering of 16 S amplicon sequences, the final OTU table consisted of 3 642 446 total reads representing 762 97% OTUs. The minimum and maximum number of reads per sample (n=99) were 21 983 and 83 157, respectively (mean=36 972), and read counts did not differ among age classes (F9,89=1.3, P=0.26). Of the top 50 most abundant OTUs detected across pitcher samples, the majority belonged to families Bacteroidetes (Supplementary Figure 3), Firmicutes (Supplementary Figure 4) and Proteobacteria (Supplementary Figure 5). As hypothesized in Figure 1a, bacterial alpha diversities (Shannon’s H’) peaked at the end of the first growing season and experienced a slight decrease after day 88 (Figure 2b), whereas phylogenetic diversity increased over the entire study period (Supplementary Figure 2). Taxonomic richness was highly correlated with phylogenetic diversity (Pearson’s r=0.96)—increasing over time with the greatest variation among the 365-day samples. In contrast with the prediction in Figure 1c, however, community composition tended to converge (that is, beta diversity decreased) during the course of the first growing season, and diverge again during the start of the second growing season, according to both taxonomic (Figure 2c) and phylogenetic (Supplementary Figure 2) dissimilarity metrics. Furthermore, permutational ANOVA on Jensen-Shannon and UniFrac distances revealed a structuring of pitcher bacterial communities by age class (Supplementary Table 1) and parallel successional trajectories between years (Figure 3a, Supplementary Figure 6).

Figure 3.

Figure 3

Open in a new tab

Principal coordinate (PCoA) plots for (a) Jensen-Shannon distances between samples, demonstrating convergence and approximately parallel successional trajectories in between-population community structures over time, and (b) Jaccard distances between Biolog plates for communities of different ages, demonstrating the convergence of metabolic profiles in mid-successional pitcher leaves and overlapping metabolic profiles for young and senescing leaves. The percentages of variance explained by the principal coordinates are displayed on each axis. Points denote yearly centroid values±s.e.m.

A subset of OTUs experienced particularly strong temporal turnover (Figure 4). These taxa fell primarily into the phyla Proteobacteria (37 OTUs), Bacteroidetes (16 OTUs) and Firmicutes (14 OTUs). Using these OTUs to train a random forest classifier to predict the pitcher community’s age resulted in a high classification accuracy for withheld data (observed vs predicted R2=0.80). Likewise, a random forest trained on 2013 data was successful at predicting the ages of samples collected from the independent 2014 population (R2=0.75) (Supplementary Figure 8), implying that observed community trajectories are parallel and generalizable between individuals and populations.

Figure 4.

Figure 4

Open in a new tab

Abundance-weighted heat map of 97% OTUs that experienced significant (P<0.01) 8-fold or greater turnover between time points for the (a) 2013 Blackhawk Creek and (b) 2014 Butterfly Valley study populations. Tick marks on X-axis denote individual pitcher samples. OTUs are labeled by family and ordered based on the community age in which they were first detected regardless of year. Random forest models trained on OTU abundances from 2013 were able to predict the ages of 2014 samples with 75% accuracy (Supplementary Figure 7).

Temporal trends in the functional attributes of pitcher microbiota

Assays of pitcher communities’ carbon substrate use patterns mirrored trends in taxonomic and phylogenetic alpha and beta diversities—namely, early and late-stage pitcher communities both metabolized significantly fewer carbon substrates than did 55-day communities (Figure 5a). Furthermore, 11-day and 365-day pitchers’ substrate profiles were much more variable than and clustered apart from the 55-day samples. (Figure 3b; Figure 5b).

Figure 5.

Figure 5

Open in a new tab

(a) Mid-successional pitcher communities were capable of metabolizing significantly more Biolog GN2 plate C-substrates than were early- and late-stage communities. (b) Mid-successional pitcher communities were much more similar to one another in terms of their carbon metabolic profiles than were early- and late-stage pitchers.

A PCA plot of samples’ reconstructed metagenomes predicted pitcher samples to separate by age, with the greatest distances between the 11-day and 365-day communities (Supplementary Figure 8). The average number of rRNA gene copies per taxon was predicted to be greater in 11-day pitchers than in any other age class (Supplementary Figure 9). This trend was also observed in the relative abundances of a number of other predicted KEGG pathways, such as flagellar assembly, motility, chemotaxis, and ABC transporters (Supplementary Figure 10). Conversely, a variety of metabolic pathways were predicted to increase over time (Supplementary Figure 11). Likewise, the abundances of genes involved in nitrogen cycling (deamination, nitrogen mineralization, denitrification, and nitrogen fixation) were also predicted to increase over a pitcher leaf’s lifespan (Supplementary Figures 12–15).

Linking community dynamics and ecosystem properties

Prey decomposition was unimodal over leaves’ lifespans, peaking at 44–88 days (Figure 6a). This increased decomposition, however, did not herald similar temporal differences in common-garden community respiration rates, although there was still a positive, non-significant unimodal trend in mean respiration rates over time (Supplementary Figure 2). Multinomial logit models predicted bacterial diversity, bacterial abundance, and midge abundance to positively influence a pitcher’s probability of having a higher decomposition score (Figures 6b and c; Table 1). Leaf nitrogen uptake efficiency also increased during the first growing season and subsequently declined at the start of year 2 (Figure 6d), and was found to be positively associated with decomposition extent and leaf dry mass (Figure 6e; Table 1). Additionally, there was a weak but significant positive correlation between pitcher samples’ JSD/unweighted UniFrac distances and their Euclidean distances in nitrogen uptake efficiencies (JSD Mantel r=0.08, P<0.05; UniFrac Mantel r=0.10, P<0.05). Finally, in contrast to natural pitcher samples collected in 2014, the nitrogen uptake efficiencies of experimentally homogenized pitcher food webs did not differ between leaf age classes (F2,12=0.98, P=0.40) (Figure 7).

Figure 6.

Figure 6

Open in a new tab

Trends in ecosystem properties during succession. (a) The frequencies of decomposition classes for pitchers of different ages. Square size is proportional to relative frequency of a particular decomposition category for that age class. χ2 is the likelihood ratio test statistic for the effect of pitcher age on the fit of a multinomial logit distribution to predict decomposition categories. (b, c) The probabilities of observing high decomposition rates increases with both bacterial diversity and bacterial biomass. Curves represent fitted proabilities of multinomial logit models, and individual curves can be interpreted as logistic regression fits for each decomposition category. (d) Pitcher leaves’ nitrogen uptake efficiencies change over time, and are significantly lower in late-stage pitcher leaves. Points denote mean values±s.e.m. (e) The extent of prey decomposition is positively associated with the percentage of prey-derived nitrogen found in the host leaf’s foliar tissue. The dashed line denotes the best-fit linear model±95% CI.

Table 1. Model selection results of multinomial logit and linear regression models for decomposition category and nitrogen uptake efficiency, respectively.

Decomposition category
 Nitrogen uptake efficiency
Predictor variables ΔAICc Pseudo-R2 Predictor variables ΔAICc R2
Community agea (A) 32 0.71 Community agea (A) 418 0.17
Bacterial abundance (B) 23 0.37 Bacterial abundance (B) 420 0.04
Bacterial diversity (D) 26 0.34 Bacterial diversity (D) 430 0.05
Bacterivore richness (R) 61 0.05 Log midge abundance (M) 436 0.00
Log midge abundance (M) 45 0.20 Log mite abundance (N) 436 0.01
Log mite abundance (N) 60 0.05 Leaf dry mass (P) 410 0.23
B+D 0 0.56 Decomposition category (C) 14 0.16
B+D+M 3 0.59 A+C+P 4 0.37
A+B+D+M 24 0.82 A+B+D+P+C 0 0.34
1 (intercept-only null) 55 0.00 1 (intercept-only null) 434 0.00
Open in a new tab

Bolded values indicate the best-performing models based on AICc and R2 statistics. AICc values falling within 9 units of the top model were considered equally parsimonious.

a

Age covariate was modeled as quadratic.

Figure 7.

Figure 7

Open in a new tab

Homogenizing the food webs of 11, 55 and 365 day pitchers and placing them back into the plants removes the significant differences observed in natural pitcher communities of the same ages. Letters above the groups represent the within-treatment contrasts. Points denote mean values±s.e.m. (n=5).

Discussion

As predicted, community diversity and biomass were positively associated with rates of prey decomposition, and the extent of decomposition was positively associated with the fraction of prey-derived nitrogen removed from the food web by the host leaf. In concert, these results imply that the services these digestive communities provide their hosts are time-dependent—highlighting important, general linkages between the temporal dynamics of communities and rates of ecosystem or host function.

Temporal patterns in community composition

The logistic-like accumulation of both bacterial diversity and biomass in developing pitcher leaves aligns with both predictions from succession models (Figure 1b) and time series of animal gut communities (Koenig et al., 2011; Jemielita et al., 2014). However, it is important to note that both bacterial and midge abundances decreased over the winter—likely in response to the cessation of prey capture. A unimodal or monotonic increase in diversity over time is anticipated for open systems experiencing high rates of immigration and low rates of extinction, which is the likely state of pitcher leaves during their first growing season. Once leaves cease to produce prey attractants, prey capture becomes more stochastic (Wolfe, 1981). Because of this, bacterial communities may experience extinctions under diminishing resource levels or continue to accumulate diversity if prey capture continues to occur. This may explain the increased variation in diversity among year-old leaves.

Because pitcher plant leaves are similar in habitat structure and resource composition at a particular point in time, it is not surprising that bacterial communities converged in composition over the first growing season. This convergence can be attributed to common selection pressures acting on a shared pool of immigrants, which would serve to homogenize communities (Vellend, 2016). This convergence is supported by the converging carbon substrate and OTU profiles of pitcher communities from two different populations. Successional convergence has also been documented in non-bacterial communities from the pitcher plant Sarracenia purpurea (Miller and terHorst, 2012), other phyllosphere bacterial communities (Copeland et al., 2015), the human gut (Palmer et al., 2007), and more generally, across a variety of terrestrial (Christensen and Peet, 1984) and aquatic ecosystems (Moorhead et al., 1998).

Contrasting with this pattern, year-old leaves contained higher microbial beta diversities than those observed in preceding time points, implying communities diverged over the winter. This is likely the consequence of stochastic prey capture amplifying differences in leaves’ ratios of labile to recalcitrant metabolic substrates. If this ratio constitutes a reasonably strong selection gradient, then this heterogeneity should drive divergence among communities (Eisenhauer et al., 2013; Dini-Andreote et al., 2015). Alternatively, stochastic drift can drive community divergence when the number of individuals is small (Orrock and Watling, 2010; Vellend, 2016). In Darlingtonia leaves, however, drift is likely minimal, since bacterial population sizes are probably too large to be influenced by demographic stochasticity.

Temporal variation in propagule supply can also lead to community divergence (Evans et al., 2017). This might occur when a fraction of ageing leaves, whose communities had previously been homogenized by a sustained input from a common microbial pool, suddenly experience a more stochastic supply of immigrants. If the species comprising the common immigrant pool also vary over time, then discontinuous, stochastic prey input could drive divergence in communities in the absence of drift and selection effects. A 55-year study of old-field communities observed similar patterns of convergence giving way to divergence driven by dispersal limitation (Meiners et al., 2015). Nonlinear temporal trends in beta diversity have also been identified in host-associated and groundwater microbial communities (Marino et al., 2014; Zhou et al., 2014), though the processes governing these patterns remain vague. Pitcher microbial communities offer a tractable system in which to experimentally assess the relative influences of deterministic vs stochastic dispersal on beta diversity.

Temporal trends in communities’ functional attributes

Leaf communities’ carbon metabolic profiles had temporal patterns similar to OTU beta diversity, implicating a link between community composition and metabolic functioning. However, microbial community sequences were not generated from the leaves used for Biolog assays, prohibiting a direct test of this hypothesis. Many of the genes predicted to be enriched in young pitchers (ribosomal RNA copy number, chemotaxis/motility genes) have been linked to a taxon’s responsiveness to unpredictable nutrient conditions (Klappenbach et al., 2000; Livermore et al., 2014; Nemergut et al., 2015). These predictions are in accordance with successional tolerance and inhibition models, wherein ruderal, fast-responders are eventually joined or outcompeted by more growth-efficient forms (Connell and Slatyer, 1977; Huston and Smith, 1987; Tilman, 1990).

Metabolic pathways contributing to amino acid demamination and N mineralization were predicted to be enriched during mid-succession—a pattern also detected during microbial succession on decomposing corpses (Metcalf et al., 2016). Similar successional increases in metabolic genes have been documented in host-associated (Koenig et al., 2011) and aquatic (Teeling et al., 2012) bacterial communities. In concert with the community metabolic assays, these findings demonstrate, in principle, how bacterial communities’ taxonomic and functional profiles can undergo predictable changes over a host’s lifespan in accordance with predictions derived from succession models. The next step is to relate these community changes to the services they provide the host organism.

Linking community properties to host functioning

In agreement with succession hypotheses (Figure 1d), detrital processing rates by the pitcher leaf communities varied over time, and were positively associated with detritivore abundances (bacteria, midge larvae) and bacterial diversity. Loreau (2001) reasoned that microbial diversity would enhance decomposition only if the number of organic compounds able to be metabolized by the community increased with alpha diversity. This prediction is supported by observations of peak bacterial diversity coinciding with peak carbon metabolic diversity during mid-succession (ca. 55 days). To date, the few studies to investigate microbial diversity and decomposition rates in situ have arrived at conflicting results (Hättenschwiler et al., 2011) but a positive relationship is common in the few experimental tests using bacteria (Nielsen et al., 2011), including in a lab experiment using bacterial isolates from the same Darlingtonia population studied here (Armitage, 2016). More generally, microbial community composition is anticipated to set ecosystem process rates (Figure 1e)—especially when the effects of environmental variation are minimal (Graham et al., 2016).

From a host plant’s perspective, decomposition by its commensal biota should set limits on its rate of N sequestration. In Darlingtonia, the state of fly digestion explained a some of the variance in N-uptake efficiency, though there was still a large amount of unexplained variance to account for. More convincingly, a follow-up experiment failed to detect the same mid-succession peak in N-uptake efficiencies among pitcher leaves containing experimentally homogenized bacterial communities. The related pitcher plant Sarracenia purpurea also relies heavily on its bacterial community for nitrogen processing (Butler et al., 2008). Furthermore, microbial community composition is an important determinant of nitrogen mineralization rates in soil (Balser and Firestone, 2005; Strickland et al., 2009), and changes in N mineralization can track microbial community change over time, independent of environmental variation (Balser and Firestone, 2005). In concert, these results highlight the potential for the pitcher microbial communities to mediate N transfer from prey to host—a function critical to the fitness of a host plant adapted to life in nitrogen-poor soils.

Contrary to predictions from succession models (Vitousek and Reiners, 1975; Finn, 1982; Loreau, 1998), maximal rates of N loss from the Darlingtonia food web occurred during periods of high (rather than low) standing biomass. This mismatch may be explained by differences between donor-controlled food webs, which receive pulses of bioavailable N at regular intervals, and primary producer-controlled food webs, in which the N pool is slowly renewed in situ and quickly immobilized (Fierer et al., 2010). As a consequence, donor-controlled food webs may not experience strong competitive pressure to sequester growth-limiting nutrients. This may be particularly true in Darlingtonia and other digestive communities for two reasons. First, rapid bacterial turnover (for example, via viral lysis & protozoan grazing) serves to increase the concentration of bioavailable N. Second, pitcher leaves’ continuous accumulation of low C:N detritus (relative to plant-based food webs) may buffer the food web from a loss of nitrogen to the host plant.

Succession or seasonality?

Because study leaves belonged to the same cohort, their temporal dynamics may reflect the effects of seasonal forcing rather than succession. Although winter temperatures drive the plants into a state of dormancy, their leaves persist, and I have observed active populations of mites, midges and bacteria in pitcher leaves underneath snow cover, suggesting that the food web still functions during the winter months. Furthermore, these brief cold periods are unlikely to have caused strong population bottlenecks or extinctions, given the large bacterial biomasses observed across pitcher leaves. Seasonal forcing should cause community composition to be cyclical over an annual cycle, yet communities collected from 11 and 365-day samples on the same day were strongly dissimilar, implying that under nearly identical external environmental conditions, communities show measurable age-related differences—an observation in line with previous studies (Thompson et al., 1993; Redford and Fierer, 2009; Williams et al., 2013; Metcalf et al., 2016).

Cross-system considerations

It is now recognized that community and ecosystem dynamics are shaped by unique combinations of disturbances, competition, and dispersal (Meiners et al., 2015). And although succession is most frequently defined in terms of species turnover, it is reasonable to redefine it as the change in average trait values or gene frequencies within a community. Such change could influence the functioning of the host plant if, for instance, selection favored a more efficient processing or storage of nitrogen by commensal organisms. The potential for rapid evolutionary change to influence ecosystem properties has been documented (Harmon et al., 2009), yet theory integrating ecosystem development and evolution is scarce (Loreau, 1998). In doing so, care must be taken to avoid ascribing adaptive properties to ecosystems (that is, treating ecosystems as ‘super-organisms’) (Odum, 1969). However, because many host-associated systems serve functions critical to their hosts’ fitnesses, they may be expected to more closely align with Odum’s controversial predictions for increasing stability and productivity. Tests of these predictions (Beaver, 1985; Neutel et al., 2007) using existing quantitative frameworks (Finn, 1982; DeAngelis, 1992; Loreau, 1998) would be difficult but valuable contributions toward a unified theory of communities and ecosystems.

Conclusions

By combining a 15N stable isotope pulse-chase experiment with observations of community dynamics, I have confirmed a number of successional hypotheses in natural, host-associated microbial digestive communities. In particular, my data support and extend the hypotheses of parallel community trajectories and mid-successional peaks in functional and taxonomic diversity to host-associated bacterial communities. In concert, these results represent a step towards integrating host-associated microbial communities into classical conceptual models of ecosystem development and demonstrate a coupling of community dynamics and host functioning. Looking ahead, more theoretical and experimental work is needed before we can identify definitive links between community dynamics and host functioning, and I believe that the continued experimental use of replicated, natural host-associated communities offers a productive path forward.

Acknowledgments

I thank Anna Petrosky, Ramon Leon, & Stefani Brandt for assistance with data collection. Ellen Simms, Todd Dawson, and the UC Berkeley Forestry Camp provided facilities and equipment. I thank Stuart Jones, Mary Firestone, Mary Power & Wayne Sousa for critical feedback. Field collection permits were provided by Jim Belsher-Howe (USFS). Data are publically available on the MG-RAST server under project ID mgp14344. Funding was provided by NSF DEB-1406524 & an NSF GRFP.

Author contributions

DWA conceived this work, performed data collection and analysis, and wrote the manuscript.

Footnotes

Supplementary Information accompanies this paper on The ISME Journal website (http://www.nature.com/ismej)

The author declares no conflict of interest.

Supplementary Material

Supplementary Information
Click here for additional data file. (5.4MB, docx)

References

  1. Agresti A. (2013) Categorical Data Analysis, 3rd ed. John Wiley & Sons, Inc.: Hoboken, NJ, USA. [Google Scholar]
  2. Alday JG, Marrs RH, Martínez-Ruiz C. (2011). Vegetation convergence during early succession on coal wastes: a 6-year permanent plot study. J Veg Sci 22: 1072–1083. [Google Scholar]
  3. Anderson MJ. (2001). A new method for non-parametric multivariate analysis of variance. Austral Ecol 26: 32–46. [Google Scholar]
  4. Armitage DW. (2016). Time-variant species pools shape competitive dynamics and biodiversity–ecosystem function relationships. Proc R Soc B 283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Auclair AN, Goff FG. (1971). Diversity relations of upland forests in the Western Great Lakes area. Am Nat 105: 499–528. [Google Scholar]
  6. Bäckhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon JI. (2005). Host-bacterial mutualism in the human intestine. Science 307: 1915–1920. [DOI] [PubMed] [Google Scholar]
  7. Balser TC, Firestone MK. (2005). Linking microbial community composition and soil processes in a California annual grassland and mixed-conifer forest. Biogeochemistry 73: 395–415. [Google Scholar]
  8. Beaver RA. (1985). Geographical variation in food web structure in Nepenthes pitcher plants. Ecol Entomol 10: 241–248. [Google Scholar]
  9. Bell T, Newman JA, Silverman BW, Turner SL, Lilley AK. (2005). The contribution of species richness and composition to bacterial services. Nature 436: 1157–1160. [DOI] [PubMed] [Google Scholar]
  10. Burnham KP, Anderson DR. (2003) Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Springer Science & Business Media: Berlin, Germany. [Google Scholar]
  11. Butler JL, Gotelli NJ, Ellison AM. (2008). Linking the brown and green: nutrient transformation and fate in the Sarracenia microecosystem. Ecology 89: 898–904. [DOI] [PubMed] [Google Scholar]
  12. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK et al. (2010). QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7: 335–336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cardinale BJ, Wright JP, Cadotte MW, Carroll IT, Hector A, Srivastava DS et al. (2007). Impacts of plant diversity on biomass production increase through time because of species complementarity. Proc Natl Acad Sci USA 104: 18123–18128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Christensen NL, Peet RK. (1984). Convergence during secondary forest succession. J Ecol 72: 25–36. [Google Scholar]
  15. Connell JH, Slatyer RO. (1977). Mechanisms of succession in natural communities and their role in community stability and organization. Am Nat 111: 1119–1144. [Google Scholar]
  16. Copeland JK, Yuan L, Layeghifard M, Wang PW, Guttman DS. (2015). Seasonal community succession of the phyllosphere microbiome. Mol Plant Microbe Interact 28: 274–285. [DOI] [PubMed] [Google Scholar]
  17. DeAngelis DL. (1992) Dynamics of Nutrient Cycling and Food Webs. Chapman & Hall: London, UK; New York, NY, USA. [Google Scholar]
  18. Dini-Andreote F, Stegen JC, Elsas JD, van, Salles JF. (2015). Disentangling mechanisms that mediate the balance between stochastic and deterministic processes in microbial succession. Proc Natl Acad Sci USA 112: E1326–E1332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Eisenhauer N, Schulz W, Scheu S, Jousset A. (2013). Niche dimensionality links biodiversity and invasibility of microbial communities. Funct Ecol 27: 282–288. [Google Scholar]
  20. Evans S, Martiny JBH, Allison SD. (2017). Effects of dispersal and selection on stochastic assembly in microbial communities. ISME J 11: 176–185. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fierer N, Nemergut D, Knight R, Craine JM. (2010). Changes through time: integrating microorganisms into the study of succession. Res Microbiol 161: 635–642. [DOI] [PubMed] [Google Scholar]
  22. Finn JT. (1982). Ecosystem succession, nutrient cycling and output-input ratios. J Theor Biol 99: 479–489. [Google Scholar]
  23. Fisher SG, Gray LJ, Grimm NB, Busch DE. (1982). Temporal succession in a desert stream ecosystem following flash flooding. Ecol Monogr 52: 93–110. [Google Scholar]
  24. Graham EB, Knelman JE, Schindlbacher A, Siciliano S, Breulmann M, Yannarell A et al. (2016). Microbes as engines of ecosystem function: when does community structure enhance predictions of ecosystem processes? Front Microbiol 7: 214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Harmon LJ, Matthews B, Des Roches S, Chase JM, Shurin JB, Schluter D. (2009). Evolutionary diversification in stickleback affects ecosystem functioning. Nature 458: 1167–1170. [DOI] [PubMed] [Google Scholar]
  26. Hättenschwiler S, Fromin N, Barantal S. (2011). Functional diversity of terrestrial microbial decomposers and their substrates. C R Biol 334: 393–402. [DOI] [PubMed] [Google Scholar]
  27. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S et al. (2005). Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecol Monogr 75: 3–35. [Google Scholar]
  28. Huston M, Smith T. (1987). Plant succession: life history and competition. Am Nat 130: 168–198. [Google Scholar]
  29. Jemielita M, Taormina MJ, Burns AR, Hampton JS, Rolig AS, Guillemin K et al. (2014). Spatial and temporal features of the growth of a bacterial species colonizing the zebrafish gut. mBio 5: e01751–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kanehisa M, Sato Y, Kawashima M, Furumichi M, Tanabe M. (2016). KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res 44: D457–D462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Klappenbach JA, Dunbar JM, Schmidt TM. (2000). rRNA operon copy number reflects ecological strategies of bacteria. Appl Environ Microbiol 66: 1328–1333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Koenig JE, Spor A, Scalfone N, Fricker AD, Stombaugh J, Knight R et al. (2011). Succession of microbial consortia in the developing infant gut microbiome. Proc Natl Acad Sci USA 108: 4578–4585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Langille MGI, Zaneveld J, Caporaso JG, McDonald D, Knights D, Reyes JA et al. (2013). Predictive functional profiling of microbial communities using 16 S rRNA marker gene sequences. Nat Biotechnol 31: 814–821. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Livermore JA, Emrich SJ, Tan J, Jones SE. (2014). Freshwater bacterial lifestyles inferred from comparative genomics. Environ Microbiol 16: 746–758. [DOI] [PubMed] [Google Scholar]
  35. Loreau M. (1998). Ecosystem development explained by competition within and between material cycles. Proc R Soc B Biol Sci 265: 33–38. [Google Scholar]
  36. Loreau M. (2001). Microbial diversity, producer–decomposer interactions and ecosystem processes: a theoretical model. Proc R Soc Lond B Biol Sci 268: 303–309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Loucks OL. (1970). Evolution of diversity, efficiency, and community stability. Am Zool 10: 17–25. [DOI] [PubMed] [Google Scholar]
  38. Love MI, Huber W, Anders S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15: 1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Lugtenberg B, Kamilova F. (2009). Plant-growth-promoting rhizobacteria. Annu Rev Microbiol 63: 541–556. [DOI] [PubMed] [Google Scholar]
  40. Marino S, Baxter NT, Huffnagle GB, Petrosino JF, Schloss PD. (2014). Mathematical modeling of primary succession of murine intestinal microbiota. Proc Natl Acad Sci USA 111: 439–444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. McMurdie PJ, Holmes S. (2013). phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLOS ONE 8: e61217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Meiners SJ, Pickett STA, Cadenasso ML. (2015) An Integrative Approach to Successional Dynamics: Tempo and Mode of Vegetation Change. Cambridge University Press: Cambridge, UK. [Google Scholar]
  43. Metcalf JL, Xu ZZ, Weiss S, Lax S, Treuren WV, Hyde ER et al. (2016). Microbial community assembly and metabolic function during mammalian corpse decomposition. Science 351: 158–162. [DOI] [PubMed] [Google Scholar]
  44. Miller TE, terHorst CP. (2012). Testing successional hypotheses of stability, heterogeneity, and diversity in pitcher-plant inquiline communities. Oecologia 170: 243–251. [DOI] [PubMed] [Google Scholar]
  45. Moorhead DL, Hall DL, Willig MR. (1998). Succession of macroinvertebrates in playas of the Southern High Plains, USA. J North Am Benthol Soc 17: 430–442. [Google Scholar]
  46. Nemergut DR, Knelman JE, Ferrenberg S, Bilinski T, Melbourne B, Jiang L et al. (2015). Decreases in average bacterial community rRNA operon copy number during succession. ISME J 10: 1147–1156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Neutel A-M, Heesterbeek JAP, van de Koppel J, Hoenderboom G, Vos A, Kaldeway C et al. (2007). Reconciling complexity with stability in naturally assembling food webs. Nature 449: 599–602. [DOI] [PubMed] [Google Scholar]
  48. Nielsen UN, Ayres E, Wall DH, Bardgett RD. (2011). Soil biodiversity and carbon cycling: a review and synthesis of studies examining diversity–function relationships. Eur J Soil Sci 62: 105–116. [Google Scholar]
  49. Odum EP. (1969). The strategy of ecosystem development. Science 164: 262–270. [DOI] [PubMed] [Google Scholar]
  50. Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O’Hara RB et al. (2015), vegan: Community Ecology Package.
  51. Orrock JL, Watling JI. (2010). Local community size mediates ecological drift and competition in metacommunities. Proc R Soc Lond B Biol Sci 277: 2185–2191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Palmer C, Bik EM, DiGiulio DB, Relman DA, Brown PO. (2007). Development of the human infant intestinal microbiota. PLOS Biol 5: e177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Peet RK, Christensen NL. (1988) Changes in species diversity during secondary forest succession on the North Carolina Piedmont. In: During HJ, Werger MJA, Williams JH (eds). Diversity and Pattern in Plant Communities. SPB Academic Publishing: The Hague, The Netherlands, pp 233–245. [Google Scholar]
  54. R Development Core Team. (2015) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing: Vienna, Austria. [Google Scholar]
  55. Redford AJ, Fierer N. (2009). Bacterial succession on the leaf surface: a novel system for studying successional dynamics. Microb Ecol 58: 189–198. [DOI] [PubMed] [Google Scholar]
  56. van Ruijven J, Berendse F. (2005). Diversity-productivity relationships: initial effects, long-term patterns, and underlying mechanisms. Proc Natl Acad Sci USA 102: 695–700. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Schmidt SK, Costello EK, Nemergut DR, Cleveland CC, Reed SC, Weintraub MN et al. (2007). Biogeochemical consequences of rapid microbial turnover and seasonal succession in soil. Ecology 88: 1379–1385. [DOI] [PubMed] [Google Scholar]
  58. Strickland MS, Lauber C, Fierer N, Bradford MA. (2009). Testing the functional significance of microbial community composition. Ecology 90: 441–451. [DOI] [PubMed] [Google Scholar]
  59. Teeling H, Fuchs BM, Becher D, Klockow C, Gardebrecht A, Bennke CM et al. (2012). Substrate-controlled succession of marine bacterioplankton populations induced by a phytoplankton bloom. Science 336: 608–611. [DOI] [PubMed] [Google Scholar]
  60. Thompson IP, Bailey MJ, Fenlon JS, Fermor TR, Lilley AK, Lynch JM et al. (1993). Quantitative and qualitative seasonal changes in the microbial community from the phyllosphere of sugar beet (Beta vulgaris). Plant Soil 150: 177–191. [Google Scholar]
  61. Tilman D. (1990). Constraints and tradeoffs: toward a predictive theory of competition and succession. Oikos 58: 3–15. [Google Scholar]
  62. Vellend M. (2016) The Theory of Ecological Communities. Princeton University Press: Princeton, NJ, USA. [Google Scholar]
  63. Vitousek PM, Reiners WA. (1975). Ecosystem succession and nutrient retention: a hypothesis. Bioscience 25: 376–381. [Google Scholar]
  64. Weis JJ, Cardinale BJ, Forshay KJ, Ives AR. (2007). Effects of species diversity on community biomass production change over the course of succession. Ecology 88: 929–939. [DOI] [PubMed] [Google Scholar]
  65. Williams MA, Jangid K, Shanmugam SG, Whitman WB. (2013). Bacterial communities in soil mimic patterns of vegetative succession and ecosystem climax but are resilient to change between seasons. Soil Biol Biochem 57: 749–757. [Google Scholar]
  66. Wolfe LM. (1981). Feeding behavior of a plant: differential prey capture in old and new leaves of the pitcher plant (Sarracenia purpurea. Am Midl Nat 106: 352–359. [Google Scholar]
  67. Zhou J, Deng Y, Zhang P, Xue K, Liang Y, Nostrand JDV et al. (2014). Stochasticity, succession, and environmental perturbations in a fluidic ecosystem. Proc Natl Acad Sci USA 111: E836–E845. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information
Click here for additional data file. (5.4MB, docx)

Articles from The ISME Journal are provided here courtesy of Oxford University Press

RESOURCES