The Effects of Host Age and Spatial Location on Bacterial Community Composition in the English Oak Tree (Quercus Robur)

Environmental Microbiology Reports (2016) 8(5), 649–658 doi:10.1111/1758-2229.12418

The effects of host age and spatial location on bacterial community composition in the English Oak tree (Quercus robur)

Meaden S.,1* Metcalf C.J.E.2,3 and Koskella B.4 poor understanding of which mechanisms shape the 1College of Life and Environmental Sciences, University plant-associated microbial community or how this might of Exeter, Penryn Campus, TR109FE, United Kingdom. in-turn influence host traits. Furthermore, although plant 2Department of Ecology and Evolutionary Biology, microbiome research has primarily focused on the below Princeton University, Princeton, USA. ground portion of the plant (the rhizosphere), knowledge 3Fogarty International Center, National Institute of of the phylloplane (the microbial composition of leaves) Health, Bethesda, Maryland, USA. is increasing (Lindow and Brandl, 2003; Vorholt, 2012), 4Department of Integrative Biology, University of demonstrating an equally important role in shaping plant California, Berkeley, 94720, USA. phenotype. Still less is known regarding the microbial composition of other organs, with distinct communities Summary reported across tissues within the host, often playing a more important role than biogeography (Ottesen et al., Drivers of bacterial community assemblages associ- 2013; Leff et al., 2014; Coleman-Derr et al., 2016). For ated with plants are diverse and include biotic fac- tree species in particular, the dermosphere (bark associ- tors, such as competitors and host traits, and abiotic ated microbial community, (Lambais et al., 2014) may factors, including environmental conditions and dis- be particularly important given that bacterial pathogens persal mechanisms. We examine the roles of spatial often invade the host through wounds in the bark (Tattar, distribution and host size, as an approximation for 2012; Misas-Villamil et al., 2013). This variation among age, in shaping the microbiome associated with tissues mirrors what is observed in other long-lived Quercus robur woody tissue using culture- hosts, including humans, where data is most abundant; independent 16S rRNA gene amplicon sequencing. In distinct bacterial communities have been isolated from addition to providing a baseline survey of the Q. different skin sites (Grice et al., 2009) and these differ- robur microbiome, we screened for the pathogen of ences appear stable over time (Costello et al., 2009). acute oak decline. Our results suggest that age is a Such variation is also likely to exist across individual predictor of bacterial community composition, dem- plant microbiomes given that they can be heritable onstrating a surprising negative correlation between (Peiffer et al., 2013), shaped by host genetics (Boden- tree age and alpha diversity. We find no signature of hausen et al., 2014; Beckers et al., 2016), and play dispersal limitation within the Wytham Woods plot functional roles that include sensitizing the plant immune sampled. Together, these results provide evidence for niche-based hypotheses of community assembly and system (Pieterse et al., 2014). the importance of tree age in bacterial community The root-associated microbiomes of healthy Arabidop- structure, as well as highlighting that caution must sis plants are arguably the best understood plant micro- be applied when diagnosing dysbiosis in a long-lived biome (Lundberg et al., 2012) with the mechanisms plant host. behind host regulation recently coming to light (Lebeis et al., 2015). However, many more non-model plant species have had their microbiomes characterized. For example, a number of studies have explored the nature Introduction of tree microbiomes, providing baseline taxonomic sur- Many lines of evidence suggest that microbes are cru- veys and assessing the drivers of community composi- cial for plant health and function (Kim et al., 2011; tion, typically contrasting host traits with climatic or Berendsen et al., 2012), and yet we have a relatively geographic variables. Many of these studies find a strong effect of host phylogeny on the bacterial commu-

Received 14 October, 2015; accepted 8 April, 2016. *For corre- nity, with a greater effect of tree species than geographic spondence. E-mail [email protected]; Tel. 15103457086. distance, even across continents (Redford et al., 2010;

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 650 S. Meaden, C. J. E. Metcalf and B. Koskella

Lambais et al., 2014). Similarly, in a tropical environment biome as they age (Tarpy et al., 2015) and in a wild bird, in Malaysia, Kim et al. (Kim et al., 2012) found a strong Rissa tridactyla, chicks harbor a greater diversity of bac- signal of host phylogeny on bacterial community compo- teria than adults (van Dongen et al., 2013). In plants, sition. Functional host traits such as growth rate and bacterial diversity can be highest on younger leaves in leaf mass have also been demonstrated as key drivers lettuces (Dees et al., 2015), however the evidence is of composition, alongside phylogeny (Kembel et al., mixed as tree-associated bacterial communities can be 2014). In contrast, Finkel et al. (Finkel et al., 2012) strongly influenced by season (Penuelas~ et al., 2012). found trees of the same species in a different desert In this study we describe and explore the bacterial locations host distinct microbial communities. Given composition of Q.robur tree cores in a well-studied UK these conflicting results across the scales examined, it forest, Wytham Woods, in order to answer three key is unclear whether phylloplane microbiomes are subject questions: firstly, what are the typical bacterial taxa to niche-based or neutral models of community associated with this Woodland site; secondly, does geo- assembly. graphic distance affect dispersal, such that there is a Specifically, the roles of dispersal and immigration, in spatial pattern of community composition and distance combination with ecological selection and drift (Vellend, between trees; and thirdly, is Q. robur host age or loca- 2010), have been the focus of a number of theoretical tion important in structuring bacterial communities. To models of community assembly, many of which are answer these questions, we first describe the tree- applicable to microbes (Sloan et al., 2006; Nemergut associated microbiota using amplicon sequencing of the et al., 2013). The niche assembly model states that the 16S rDNA gene of 64 trees. Using a long-term woodland dispersal of bacteria is unhindered by physical con- census we then assess correlations between alpha and straints, and all organisms can be found anywhere but it beta diversity and factors such as age and spatial loca- is the environment which selects for their persistence tion. Additionally, we use these data to compare the pre- (de Wit and Bouvier, 2006). Conversely, the dispersal dicted metabolic functionality and screen our dataset for assembly hypothesis states that the biodiversity we the pathogenic clade Brenneria, the causative agent of observe can largely be explained by stochastic local acute oak decline, from which the UK Q. robur popula- extinctions and dispersal-limitation, typified by the idea tion is currently experiencing an epidemic (Denman of island biogeography (Hubbell, 2001; Volkov et al., et al., 2012). This survey presents a unique opportunity 2003). Whilst this is essentially the “niche vs. neutral” to assess the practicality of high throughput sequencing debate, Fierer (Fierer, 2008) provides the nuances of in environmental monitoring. Given the critical impor- the microbial context including the much higher species tance of detecting and preventing the emergence of tree richness and evenness, and the rapidity of species turn- diseases before large-scale spread, a better understand- over typical of most bacterial communities. ing of tree microbiomes offers additional value in The English oak tree, Quercus robur, study system surveillance. provides an opportunity to test these competing hypotheses. If microbial community assembly is purely a Experimental Procedures dispersal-driven process, we would predict a positive Study System relationship between tree age and diversity, as older organisms will have experienced more colonization Wytham Woods is one of the most intensively studied events. Such a positive relationship has been demon- tree populations in Europe and undergoes extensive sur- strated for trees and their plant epiphytes and lichens veys every 2 years0020(Butt et al. 2009). As such, it (Flores-Palacios and Garcia-Franco, 2006; Johansson provides a practical system for correlating a vast number et al., 2007), but has not been shown before in tree- of ecological variables and demographic traits and has associated bacterial communities. Alternatively, if the been the source of numerous important papers (Hunter process is strictly niche-driven, older trees could repre- et al., 1997; Morecroft et al., 2003; Butt et al., 2009). sent an alternative environment to smaller trees, favor- The UK Q. robur population is suffering a number of ing proliferation of particular species but not necessarily infectious diseases, collectively known as oak decline. harboring a greater diversity. This comprises chronic oak decline, sudden oak decline As well as dispersal, host traits are likely to govern and acute oak decline (AOD) (Denman and Webber, the microbes present. Among the host factors known to 2009). Symptoms, such as stem bleeding, are strikingly influence microbial diversity, host age is often a key pre- similar, which makes misdiagnosis with Phytophthora or dictor. Data from humans suggests diversity consistently bupestrid beetles possible. A number of bacterial spe- increases with age from birth across populations (Yatsu- cies from the Brenneria genus have been isolated from nenko et al., 2012). In insects, honey bee queens Q. robur trees suffering from AOD and it is likely that undergo massive compositional shifts in their micro- this species is the causal agent (Denman et al., 2012).

1.00

0.75

OTU.ID Bacteria;__Acidobacteria;__Acidobacteriales Bacteria;__Actinobacteria;__Frankiales Bacteria;__Actinobacteria;__Micrococcales

0.50 Bacteria;__Actinobacteria;__Pseudonocardiales Bacteria;__Actinobacteria;__Thermoleophilia Bacteria;__Proteobacteria;__Alphaproteobacteria Bacteria;__Proteobacteria;__Betaproteobacteria Relative Abundance Relative Bacteria;__Proteobacteria;__Gammaproteobacteria Unassigned;Other;Other 0.25

0.00 SOL5766 SOL5749 SOL5808 SOL5823 SOL5793 SOL5777 SOL5789 SOL5799 SOL5748 SOL5798 Bark Sample (Ascending Tree Age / Size)

Fig. 1. Randomly selected barplots showing the nine most common bacterial taxa ordered by ascending tree age (from left to right). Alpha diversity decreases in older trees and there is much variation in beta-diversity. Only those taxa with a mean relative abundance greater than 2.5% across the entire dataset were retained for the figure for clarity.

The disease has not yet been reported in Wytham diameter and age, reinforcing the view that DBH is a Woods (Kirby et al., 2014) so the absence of Brenneria good proxy for age (SI, Figure 1). species on healthy trees would buttress the existing evi- Core tissue samples were obtained using the Trephor dence that Brenneria is the primary pathogen. tool (Rossi et al., 2006), allowing for three small (approxi- mately 3 cm) microcore samples to be taken at breast height at three separate sites (North, Southwest, and Site sampling Southeast). The tool was sterilized and wiped thoroughly using 70% ethanol in between each sample extraction. We sampled 64 Q. robur trees in a single hectare, col- Samples were flash frozen in the field for transportation lecting 192 samples in over 3 days in September, 2013. back to the laboratory. Upon return to the laboratory, Tree diameter at breast height (DBH) was recorded as a samples were homogenized using a Fast-Prep 24 instru- proxy for tree age. This method is endorsed by the UK ment (MP Biomedicals) for five minutes with the addition forestry commission as a non-destructive mode of esti- of two 0.5cm steel beads. Total DNA was then extracted mating tree age (Commission, 1998). Whilst compari- from the resulting homogenate using a Qiagen DNeasy sons among trees at different sites due to crowding may Plant Mini Kit, following the protocol provided. For ampli- be inaccurate, comparisons of the same species at the fication of the V4 region of the 16S rDNA gene, the uni- same site provides reliable estimates of tree age. Fur- versal primer set GTGCCAGCMGCCGCGGTAA (5’ — ther, using trees from our dataset that had known plant- 3’) and GGACTACHVGGGTWTCTAAT (50 —30)was ing dates, we observe a linear relationship between used.

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658 652 S. Meaden, C. J. E. Metcalf and B. Koskella

Brenneria amplification using the packages ‘vegan’ (Oksanen et al., 2016) and ‘cluster’ (Maechler et al., 2015). To demonstrate that the primers used in our study could amplify Brenneria goodwinii we cultured strain 931-23 (provided by R. Jackson, University of Reading) and Results chose a random subset of 5 of the primers used to Baseline survey of the Quercus robur microbiome amplify the V4 region to test for amplification in these positive controls. The most abundant bacterial class observed within our samples was the alphaproteobacteria, with a mean rela- Bioinformatic analysis tive abundance of 26% (1/2 12 S.D.), followed by the thermoleophilia with 22% (1/2 15 S.D.), and the beta- Illumina MiSeq 250bp paired-end reads were demulti- proteobacteria, contributing a mean of 13% (1/2 plexed and de-barcoded at the sequencing centre 15 S.D.). Overall, acidobacteria, actinobacteria and pro- (Source Biosciences, Oxford). Sequences have been teobacteria were the three most abundant phyla, making deposited in the NCBI short-read archive (accession PRJNA 298668). Quality filtering of reads was con- up over 80% of OTUs (Figure 1). ducted using the Qiime (1.9.0) pipeline (Caporaso et al., Age related decline in microbial diversity 2010). Reads were joined and filtered with the default settings (Bokulich et al., 2013). Briefly, a maximum of 3 We identified a weak negative correlation between tree consecutive low quality base calls was allowed before size and species richness (using observed OTUs) when truncating the read, phred-score threshold was set at 30 controlling for uneven sampling of individual trees (GLM, 2 (which provides a 99.9% accuracy of base call), 75% of F1,8754.13, p50.0453, pseudo R 50.048) (Figure 2.). the read was required to consist of high-quality, consec- Observed OTU count was used as the measure of species utive base call and all reads with N character base calls richness, however the result was non-significant when were dropped. Open reference OTU picking was con- Faith’s phylogenetic distance or Chao 1 estimator (Chao ducted using the Uclust algorithm and the Silva 111 16S et al., 2004) was used (p50.12 and 0.16 respectively). rDNA database at the 97% identity level (Pruesse et al., There was no effect of sample orientation (cardinal direc- 2007; Edgar, 2010). Chimera removal was performed tion), and this factor was therefore excluded from the with Chimera Slayer (Haas et al., 2011); OTUs present model during stepwise model simplification. Interestingly, at abundances less than 0.005% of the dataset were these correlations strengthened when sample size was removed as were OTUs observed in only a single increased to 110 samples by using a lower rarefaction instance, as both are known to inflate diversity estimates (Bokulich et al., 2013). Mitochondrial and chloroplast sequences were also removed. This left a remaining dataset with a total of 1013881 sequences spread across 115 samples, containing a median count of 1830 sequences per sample (mean 8816, length: 251.8 bp). Using the same trimmed sequence files, open-reference OTU picking was performed against the GreenGenes (1.5) (DeSantis et al., 2006) database (as required by Picrust) using the uclust algorithm implemented in Qiime (1.9). Functional predictions of observed taxa was made using the Picrust program (Langille et al., 2013) using the Kegg orthology database (Kanehisa et al., 2014).

Statistical analysis

Rarefaction was performed for diversity analyses to a depth of 500 sequences per sample. Whilst this is relatively low for microbiome studies, we aimed to maintain high levels of biological replication at the cost of sampling depth within individual samples. Pseudo R2 values were calculated using the residual and null deviance from model outputs as described in Faraway (2006). Fig. 2. Age based decline in species richness based on species richness, as measured by observed OTUs, following rarefaction to UniFrac scores were generated in Qiime and statistical a depth of 500 sequences per sample. GLM, F1,875 4.13, analyses were performed in R (R Core Team, 2015) P 5 0.0453. Intercept 5142, slope 520.045.

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658 Age-related decline in bacterial diversity of Oak trees 653 depth (100 sequences, data not shown). A similar result Functional predictions was mirrored by beta-diversity, where tree size was a sig- In order to predict how the function of communities nificant predictor of microbial community composition using associated with our Q. robur trees changed as they both abundance weighted UniFrac scores (PERMANOVA, aged we created a predicted metagenome using the Pic- F 54.63, p50.0036, R250.052, permutations59999) 1,87 rust program (Langille et al., 2013). However, we found and unweighted UniFrac scores (PERMANOVA, no correlation between any of the predicted individual F 52.93, p50.027, R250.033, permutations59999) 1,87 genes or functional pathways associated with our when tree ID was controlled for. observed microbiome and tree size, perhaps indicating high functional redundancy of the more diverse microbiota of smaller trees. Taxa correlations Assessing spatial patterns To investigate changes in composition further we per- Finally, to look for patterns of biogeography, or dispersal formed Spearman rank correlations against tree size for limitation, we performed Mantel correlations between a each OTU in the dataset, and found no significant asso- spatial matrix from the Euclidean distances between ciations following correction for multiple testing. To fur- trees and the UniFrac scores that measure bacterial ther assess whether there were higher taxonomic level community composition. A correlation would be indica- associations between specific bacterial clades and tree tive that the spatial distribution of trees does indeed size we selected the three most abundant phyla. Collec- affect the bacterial composition of the community. There tively, the Proteobacteria, Actinobacteria and Acidobac- was no effect of abundance for either weighted (Mantel teria made up over 80% of our sequences. We found a r 5 0.0009, p50.47, permutations59999) or unweighted significant decrease in the relative abundance of Proteo- UniFrac scores (r50.002, p50.46, permutations59999), bacteria with tree size (Kendall’s rank correlation, suggesting an absence of dispersal limitation. s520.22, z523.39, p50.0007), a significant increase in the relative abundance of Actinobacteria (s50.19, Brenneria z53.04, p50.0023) and a non-significant decrease in Reassuringly, we found no sequences identified as Bren- Acidobacteria (s520.14, z522.13, p50.033) following neria in our dataset (prior to rarefaction), despite con- Bonferroni correction (Figure 3). firming that all our tested primers could successfully amplify this species following culture in vitro.

Discussion

Our study of the bacterial microbiomes of 64 English oak trees (Quercus robur) in a single woodland provides a number of insights into the drivers of bacterial community structure and dispersal (Figure 4). Firstly, our census of the microbiome of Q. robur tissue is consistent with a previous report that found the same 3 most domi- nant phyla in the roots of oak trees: Actinobacter, Pro- teobacteria and Acidobacter (Uroz et al., 2010). The high abundance of Acidobacter is also consistent with other culture-independent studies of the phyllospheric microbiota from tropical trees (Kim et al., 2012). By comparing tree size with species richness, we found no sign of an increase in bacterial diversity as trees age. This is of particular interest as it suggests Fig. 3. Taxa-specific correlations with Oak Tree age. The relative factors other than dispersal affect microbiome structure, abundance of Actinobacteria increases with tree age (triangles, as would be expected by an increase in microbial diver- solid line, intercept 5 0.25, slope 5 0.00030; Kendall’s rank correlation, s 5 0.19, z 5 3.04, P 5 0.0023), whilst the relative abundance sity with growth as a result of species accumulation. of Proteobacteria declines (crosses, dashed line, intercept 5 0.56, When observed OTUs was used as the measure of slope 520.00026; s 520.22, z 523.39, P 5 0.0007). The relative alpha diversity we found a weak but significant decline abundance of Acidobacteria also declines however this is non- significant after Bonferroni correction (s 520.14, z 522.13, in species richness with tree age. Furthermore, negative P 5 0.033). correlations between tree age and species richness

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658 654 S. Meaden, C. J. E. Metcalf and B. Koskella

Fig. 4. Conceptual diagram of potential drivers of bacterial community composition in our Oak tree system. Communities may be seeded from wind and rain driven dispersal, or colonize the plant directly from the soil during growth. Following initial colonization, the microbes must survive, and potentially thrive, in the observed niche. The niche is likely to be dictated by, among others, competition for host resources, predation and environmental conditions. were significant when the sample size was increased by function of dispersal, such that communities are reducing rarefaction depth (and therefore excluding assembled by stochastic dispersal events and local fewer samples). Detecting subtle changes in species extinctions, we would find a correlation between spatial diversity require maximal statistical power, and there is distance among trees and community dissimilarity clearly a trade-off between sampling depth and statisti- scores (beta diversity). Conversely, if microbes have cal power. Exploring this trade-off in regard to microbial unlimited dispersal within the forest, as is often community sampling clearly warrants further study as assumed, one would expect no correlation with beta alternative approaches have yet to be widely adopted diversity. Our results suggest that latter models are most (McMurdie and Holmes, 2014). Moreover, quantifying informative, whereby we find no signature of dispersal the shape of the age-diversity relationship through the limitation (i.e. the community composition of our samples tree lifetime requires longitudinal studies to build on are not influenced by the proximity of others). There is cross-sectional studies like the data presented here. the potential for microbes to disperse at global scales One suggestion for observed age-related differences is (Morris et al., 2008), however evidence for true cosmo- variation in the chemical and physiological state of the politan distribution has been mixed to date (Caporaso host tissue (van Dongen et al., 2013) and this could be et al., 2012; Finkel et al., 2012; Sul et al., 2013) and, as the case between younger and older Q.robur tree demonstrated by Bell (Bell, 2010) also in Wytham tissues. Woods, microbial dispersal limitation may be more A flat or negative correlation between tree age and important over short time scales. bacterial alpha diversity contrasts the positive associa- We also found an increase in the relative abundance tion found between epiphytic plants and lichens and tree of Actinobacter and a decrease in Proteobacter and host age (Flores-Palacios and Garcia-Franco, 2006; Acidobacter (although the latter was only nearing signifi- Johansson et al., 2007) perhaps suggesting that bacte- cance) with tree size. Mechanistically, it is hard to ria are less dispersal-limited than other tree-associated ascribe functions to whole phyla as they encompass a organisms. To explore these ideas further, and based on range of morphologies, metabolic diversity and pathoge- the conflicting niche assembly and dispersal assembly nicity (Dworkin et al., 2006). The Acidobacter are, how- hypotheses (Hubbell, 2001; de Wit and Bouvier, 2006), ever, reported to be slow growing with low metabolic we predicted that if microbiome structure is purely a rates (Ward et al., 2009), sometimes referred to as k-

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658 Age-related decline in bacterial diversity of Oak trees 655 selection strategists due to their higher abundances in been demonstrated to prevent pathogen success, partic- soils with lower resource availability (Fierer et al., 2008). ularly in fungal endophytes (Arnold et al., 2003). Carbon mineralization rate can also be a good predictor Despite being present at low numbers, many species of Acidobacter soil abundance, but how well these find- could collectively play a role in microbial community ing translates to an alternative niche, such as tree cores, function. To explore this idea further we used metage- remains unknown (Fierer et al., 2008). If this were the nomic predictions based on our 16S sequences to case in our system we would expect Acidobacter and assess functional diversity. Given that we found a signifi- Proteobacteria to be inversely correlated; but we find the cant shift in the microbial composition (at the Phylum opposite. Maignien et al. (Maignien et al., 2014) have level) with tree age, we expected to find a similar effect also suggested that phyllosphere communities are first of functional traits. We found no such trend, as no indi- colonized by r-strategists (such as Acinetobacter and vidual genes or functional pathways were over or under Pseudomonas). Moreover, when multiple OTUs of the represented in older tree samples. This lack of functional same species are present in the source community, for correlation, despite a taxonomic correlation implies a example rainfall, only one becomes established in the level of redundancy in gene pathways among bacterial phyllosphere community, indicative of niche competition phyla, or lack of sensitivity in the methods used to pre- (Maignien et al., 2014). Given that the Acidobacter are dict a metagenome. If the latter is true, and the limitation consistently isolated at high relative abundances from is the quality of annotation in metagenomic databases soil it seems likely that the soil is the major contributing then ultimately, more metagenomic sequencing may not source for the interior microbiota of oak trees. Acido- yield more insight into community function. bacter have also been detected at high relative abun- A focus on Q. robur allows us to answer some impor- dances in the trunk of Gingko bilbao trees but not in the tant applied questions: A reassuring outcome of this leaves of the same trees, again suggesting soil derived analysis was that we failed to identify a single sequence from Brenneria species. The UK oak population is rather than phyllospheric dispersal (Leff et al., 2014). undergoing an epidemic of acute oak decline (AOD) and Whether this is through transport of microbes through the Brenneria clade of bacteria have been isolated from the phloem or a function of early, seedling colonization oaks experiencing the disease (Denman et al., 2012). remains undetermined. Interestingly, and as a word of Koch’s postulates have also been reported in the Span- caution, we identified the presence of Ralstonia in our ish oak (Quercus ilex) (Poza-Carrion et al., 2008). How- negative sequencing controls, which has been identified ever acute oak decline was not found to be present in by Salter et al. (Salter et al., 2014) as a common kit Wytham in 2014 (Kirby et al., 2014) and our data sup- contaminant. However this group also includes many ports that conclusion. This further strengthens the infer- plant pathogens and wouldn’t be unexpected in our envi- ence that Brenneria is a causative agent of the disease, ronmental samples, highlighting the difficulty in identify- as suggested by Denman et al. (Denman et al., 2012). ing contaminant sequences from environmental samples Our census provides a baseline of healthy microbial and the need for negative controls. flora in UK Q. robur and comparison with trees in dis- Whilst we have described a shift in bacterial commu- eased states is a crucial area for further study. Addition- nity structure with age, the correlations between specific ally, the observed differences in microbiome among taxa and age are only present at the phylum level and differently aged trees provides a caution for defining tree not at the OTU level. The variability in genomic content, microbiome health. The healthy microbiome of a young even among closely related bacteria (Perna et al., 2001; tree may well appear similar as that of a dysbiotic micro- Guidot et al., 2007), is often used to justify a lack of eco- biome of an old tree. As such, when using microbiome logical or metabolic similarity among hosts. However studies in the context of plant health, fair comparisons there is evidence for functional convergence at higher among plant demographics must be made in order to taxonomic ranks (Philippot et al., 2010), including trophic make useful diagnoses. and biogeographic differences (Fierer et al., 2008; Phil- ippot et al., 2009). One mechanism for our observation Acknowledgements of size-based differences could be that the age of a plant is the most important factor in determining its This work was supported by a Royal Society Research Grant (to CJEM and BK). SM was funded by a studentship induced defenses (Quintero and Bowers, 2011). Indeed, at the University of Exeter, BK by a NERC independent the complex interactions between host immune systems research fellowship (NE/K00879X/1), and CJEM by a and commensal bacteria are coming to light in different Royal Society Fellowship. The authors thank Dr. Rob Jack- systems (Brestoff and Artis, 2013; Franzenburg et al., son at the University of Reading for providing an isolate of 2013). For example, the presence of commensal Brenneria, Dr. Keith Kirby for providing data on tree age microbes is non-random in a tropical tree host and has and size and Dr. Konrad Paskiewicz as well as the faculty

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658 656 S. Meaden, C. J. E. Metcalf and B. Koskella and participants of the NERC-funded population genomics leafy greens: shift in composition and decrease in rich- workshop for training in bioinformatics analyses. The ness over time. Appl. Environ. Microbiol. 81: 1530–9. 18-ha Long-Term Forest Monitoring Plot is a collaborative Denman, S., Brady, C., Kirk, S., Cleenwerck, I., Venter, S., project between the University of Oxford, the Centre for Coutinho, T., and De Vos, P. (2012) Brenneria goodwinii Ecology and Hydrology, and the Smithsonian Institution sp. nov., associated with acute oak decline in the UK. Int. CTFS Forest-GEO (HSBC Climate Partnership). J. Syst. Evol. Microbiol. 62: 2451–6. Denman, S. and Webber, J. (2009) Oak declines: new definitions and new episodes in Britain. Q. J. For. 103: References 285–290. Arnold, A.E., Mejıa, L.C., Kyllo, D., Rojas, E.I., Maynard, Z., DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Robbins, N., and Herre, E.A. (2003) Fungal endophytes Brodie, E.L., Keller, K., et al. (2006) Greengenes, a limit pathogen damage in a tropical tree. Proc. Natl. chimera-checked 16S rRNA gene database and work- Acad. Sci. U. S. A. 100: 15649–54. bench compatible with ARB. Appl. Environ. Microbiol. 72: Beckers, B., Op De Beeck, M., Weyens, N., Van Acker, R., 5069–72. Van Montagu, M., Boerjan, W., and Vangronsveld, J. De Wit, R. and Bouvier, T. (2006) “Everything is everywhere, (2016) Lignin engineering in field-grown poplar trees but, the environment selects”; what did Baas Becking and affects the endosphere bacterial microbiome. Proc. Natl. Beijerinck really say? Environ. Microbiol. 8: 755–8. Acad. Sci. U. S. A. 113: 2312–7. Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.-H., Bell, T. (2010) Experimental tests of the bacterial distance- and Stackebrandt, E. eds. (2006) The Prokaryotes decay relationship. ISME J. 4: 1357–65. Springer New York, New York, NY. Berendsen, R.L., Pieterse, C.M.J., and Bakker, P.A.H.M. Edgar, R.C. (2010) Search and clustering orders of magni- (2012) The rhizosphere microbiome and plant health. tude faster than BLAST. Bioinformatics 26: 2460–1. Trends Plant Sci. 17: 478–86. Faraway, J. (2006) Extending the Linear Model in R. Taylor Bodenhausen, N., Bortfeld-Miller, M., Ackermann, M., and and Francis, Bota Roca FL. Vorholt, J.A. (2014) A synthetic community approach Fierer, N. (2008) Microbial biogeography: patterns in micro- reveals plant genotypes affecting the phyllosphere microbial diversity across space and time. In, Zengler,K. (ed), biota. PLoS Genet. 10: e1004283. Accessing Uncultivated Microorganisms: from the Envi- Bokulich, N.A., Subramanian, S., Faith, J.J., Gevers, D., ronment to Organisms and Genomes and Back. ASM Gordon, J.I., Knight, R., et al. (2013) Quality-filtering Press, Washington DC, pp. 95–115. vastly improves diversity estimates from Illumina amplicon Fierer, N., Bradford, M.A., and Jackson, R.B. (2008) Toward sequencing. Nat. Methods 10: 57–9. an ecological classification of soil bacteria. Ecology. 88: Brestoff, J.R. and Artis, D. (2013) Commensal bacteria at 1354–1364. the interface of host metabolism and the immune system. Finkel, O.M., Burch, A.Y., Elad, T., Huse, S.M., Lindow, Nat. Immunol. 14: 676–84. S.E., Post, A.F., and Belkin, S. (2012) Distance-decay Butt, N., Campbell, G., Malhi, Y., Morecroft, M., Fenn, K., relationships partially determine diversity patterns of phyl- and Thomas, M. (2009) Initial Results from Establishment losphere bacteria on Tamarix trees across the Sonoran of a Long-term Broadleaf Monitoring Plot at Wytham Desert [corrected]. Appl. Environ. Microbiol. 78: 6187–93. Woods, Oxford, UK. University of Oxford Report. Univer- Flores-Palacios, A. and Garcia-Franco, J.G. (2006) The sity of Oxford, Oxford, UK. relationship between tree size and epiphyte species rich- Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., ness: testing four different hypotheses. J. Biogeogr. 33: Bushman, F.D., Costello, E.K., et al. (2010) QIIME allows 323–330. analysis of high-throughput community sequencing data. Franzenburg, S., Walter, J., Kunzel,€ S., Wang, J., Baines, Nat. Methods 7: 335–6. J.F., Bosch, T.C.G., and Fraune, S. (2013) Distinct antimi- Caporaso,J.G.,Paszkiewicz,K.,Field,D.,Knight,R.,and crobial peptide expression determines host species- Gilbert, J.A. (2012) The Western English Channel contains specific bacterial associations. Proc. Natl. Acad. Sci. U. a persistent microbial seed bank. ISME J. 6: 1089–93. S. A. 110: E3730–8. Chao, A., Chazdon, R.L., Colwell, R.K., and Shen, T.-J. Grice, E.A., Kong, H.H., Conlan, S., Deming, C.B., Davis, (2004) A new statistical approach for assessing similarity J., Young, A.C., et al. (2009) Topographical and temporal of species composition with incidence and abundance diversity of the human skin microbiome. Science 324: data. Ecol. Lett. 8: 148–159. 1190–2. Coleman-Derr, D., Desgarennes, D., Fonseca-Garcia, C., Guidot, A., Prior, P., Schoenfeld, J., Carre`re, S., Genin, S., Gross, S., Clingenpeel, S., Woyke, T., et al. (2016) Plant and Boucher, C. (2007) Genomic structure and phylogeny compartment and biogeography affect microbiome com- of the plant pathogen Ralstonia solanacearum inferred position in cultivated and native Agave species. New Phy- from gene distribution analysis. J. Bacteriol. 189: 377–87. tol. 209: 798–811. Haas, B.J., Gevers, D., Earl, A.M., Feldgarden, M., Ward, D. V, Costello, E.K., Lauber, C.L., Hamady, M., Fierer, N., Giannoukos, G., et al. (2011) Chimeric 16S rRNA sequence Gordon, J.I., and Knight, R. (2009) Bacterial community formation and detection in Sanger and 454-pyrosequenced variation in human body habitats across space and time. PCR amplicons. Genome Res. 21: 494–504. Science 326: 1694–7. Hubbell, S.P. (2001) The Unified Neutral Theory of Biodiver- Dees, M.W., Lysøe, E., Nordskog, B., and Brurberg, M.B. sity and Biogeography (MPB-32) Princeton University (2015) Bacterial communities associated with surfaces of Press, Princeton, New Jersey, USA.

Hunter, M.D., Varley, G.C., and Gradwell, G.R. (1997) Esti- Misas-Villamil, J.C., Kolodziejek, I., Crabill, E., Kaschani, F., mating the relative roles of top-down and bottom-up Niessen, S., Shindo, T., et al. (2013) Pseudomonas syrin- forces on insect herbivore populations: A classic study gae pv. syringae uses proteasome inhibitor syringolin A revisited. Proc. Natl. Acad. Sci. 94: 9176–9181. to colonize from wound infection sites. PLoS Pathog. 9: Johansson, P., Rydin, H., and Thor, G. (2007) Tree age e1003281. relationships with epiphytic lichen diversity and lichen life Morecroft, M.D., Stokes, V.J., and Morison, J.I.L. (2003) history traits on ash in southern Sweden. Ecoscience 14: Seasonal changes in the photosynthetic capacity of can- 81–91. opy oak (Quercus robur) leaves: the impact of slow devel- Kanehisa, M., Goto, S., Sato, Y., Kawashima, M., opment on annual carbon uptake. Int. J. Biometeorol. 47: Furumichi, M., and Tanabe, M. (2014) Data, information, 221–6. knowledge and principle: back to metabolism in KEGG. Morris, C.E., Sands, D.C., Vinatzer, B.A., Glaux, C., Nucleic Acids Res. 42: D199–205. Guilbaud, C., Buffie`re, A., et al. (2008) The life history of Kembel, S.W., O’Connor, T.K., Arnold, H.K., Hubbell, S.P., the plant pathogen Pseudomonas syringae is linked to Wright, S.J., and Green, J.L. (2014) Relationships the water cycle. ISME J. 2: 321–34. between phyllosphere bacterial communities and plant Nemergut, D.R., Schmidt, S.K., Fukami, T., O’Neill, S.P., functional traits in a neotropical forest. Proc. Natl. Acad. Bilinski, T.M., Stanish, L.F., et al. (2013) Patterns and Sci. 111: 13715–13720. processes of microbial community assembly. Microbiol. Kim, M., Singh, D., Lai-Hoe, A., Go, R., Abdul Rahim, R., Mol. Biol. Rev. 77: 342–56. Ainuddin, A.N., et al. (2012) Distinctive phyllosphere bacte- Oksanen, J.F., Blanchet, F.G., Kindt, R., Legendre, P., rial communities in tropical trees. Microb. Ecol. 63: 674–81. Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Kim, Y.C., Leveau, J., McSpadden Gardener, B.B., Pierson, Henry, M.H.H. Stevens, and Wagner, H. (2016) vegan: E.A., Pierson, L.S., and Ryu, C.-M. (2011) The multifacto- Community Ecology Package. R package version 2.3-3. rial basis for plant health promotion by plant-associated Ottesen, A.R., Gonzalez Pena,~ A., White, J.R., Pettengill, bacteria. Appl. Environ. Microbiol. 77: 1548–55. J.B., Li, C., Allard, S., et al. (2013) Baseline survey of the Kirby, K.J., Bazely, D.R., Goldberg, E.A., Hall, J.E., Isted, anatomical microbial ecology of an important food plant: R., Perry, S.C., and Thomas, R.C. (2014) Changes in the Solanum lycopersicum (tomato). BMC Microbiol. 13: 114. tree and shrub layer of Wytham Woods (Southern Eng- Peiffer, J.A., Spor, A., Koren, O., Jin, Z., Tringe, S.G., land) 1974-2012: local and national trends compared. Dangl, J.L., et al. (2013) Diversity and heritability of the Forestry 87: 663–673. maize rhizosphere microbiome under field conditions. Lambais, M.R., Lucheta, A.R., and Crowley, D.E. (2014) Proc. Natl. Acad. Sci. 110: 6548–6553. Bacterial community assemblages associated with the Penuelas,~ J., Rico, L., Ogaya, R., Jump, A.S., and phyllosphere, dermosphere, and rhizosphere of tree spe- Terradas, J. (2012) Summer season and long-term cies of the Atlantic forest are host taxon dependent. drought increase the richness of bacteria and fungi in the Microb. Ecol. 68: 567–74. foliar phyllosphere of Quercus ilex in a mixed Mediterra- Langille, M.G.I., Zaneveld, J., Caporaso, J.G., McDonald, nean forest. Plant Biol. (Stuttg). 14: 565–75. D., Knights, D., Reyes, J.A., et al. (2013) Predictive func- Perna, N.T., Plunkett, G., Burland, V., Mau, B., Glasner, J.D., tional profiling of microbial communities using 16S rRNA Rose, D.J., et al. (2001) Genome sequence of enterohae- marker gene sequences. Nat. Biotechnol. 31: 814–21. morrhagic Escherichia coli O157:H7. Nature 409: 529–33. Lebeis, S.L., Paredes, S.H., Lundberg, D.S., Breakfield, N., Philippot, L., Andersson, S.G.E., Battin, T.J., Prosser, J.I., Gehring, J., McDonald, M., et al. (2015) Salicylic acid Schimel, J.P., Whitman, W.B., and Hallin, S. (2010) The modulates colonization of the root microbiome by specific ecological coherence of high bacterial taxonomic ranks. bacterial taxa. Science (80-.). 349: 860–4. Nat. Rev. Microbiol. 8: 523–9. Leff, J.W., Del Tredici, P., Friedman, W.E., and Fierer, N. Philippot, L., Bru, D., Saby, N.P.A., Cuhel, J., Arrouays, D., (2014) Spatial structuring of bacterial communities within Simek, M., and Hallin, S. (2009) Spatial patterns of bac- individual Ginkgo biloba trees. Environ. Microbiol. 17: terial taxa in nature reflect ecological traits of deep 2352–2361. branches of the 16S rRNA bacterial tree. Environ. Micro- Lindow, S.E. and Brandl, M.T. (2003) Microbiology of the biol. 11: 3096–104. phyllosphere. Appl. Environ. Microbiol. 69: 1875–83. Pieterse, C.M.J., Zamioudis, C., Berendsen, R.L., Weller, Lundberg, D.S., Lebeis, S.L., Paredes, S.H., Yourstone, S., D.M., Van Wees, S.C.M., and Bakker, P.A.H.M. (2014) Gehring, J., Malfatti, S., et al. (2012) Defining the core Induced systemic resistance by beneficial microbes. Arabidopsis thaliana root microbiome. Nature 488: 86–90. Annu. Rev. Phytopathol. 52: 347–75. Maechler, M., Rousseeuw, P., Struyf, A., Hubert, M., and Poza-Carrion, C., Aguilar, I., Gallego, F.J., Nunez-Moreno,~ Hornik, K. (2015) cluster: Cluster Analysis Basics and Y., Biosca, E.G., Gonzalez, R., et al. (2008) Brenneria Extensions. R package version 2.0.3. quercina and Serratia spp. isolated from Spanish oak Maignien, L., DeForce, E.A., Chafee, M.E., Eren, A.M., and trees: molecular characterization and development of Simmons, S.L. (2014) Ecological succession and sto- PCR primers. Plant Pathol. 57: 308–319. chastic variation in the assembly of Arabidopsis thaliana Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, phyllosphere communities. MBio 5: e00682–13. W., Peplies, J., and Glockner,€ F.O. (2007) SILVA: a com- McMurdie, P.J. and Holmes, S. Waste not, want not: Why prehensive online resource for quality checked and rarefying microbiome data is inadmissable. PLoS Com- aligned ribosomal RNA sequence data compatible with put. Biol. 10: e1003531. doi:10.1371/journal.pcbi.1003531 ARB. Nucleic Acids Res. 35: 7188–96.

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658 658 S. Meaden, C. J. E. Metcalf and B. Koskella

Quintero, C. and Bowers, M.D. (2011) Plant induced Tattar, T.A. (2012) Diseases of Shade Trees, Revised Edi- defenses depend more on plant age than previous history tion Elsevier Science, Academic Press, London, UK. of damage: implications for plant-herbivore interactions. UK Forestry Commision (1998) Estimating the Age of Large J. Chem. Ecol. 37: 992–1001. and Veteran Trees in Britain. Forestry Commision Infor- R Core Team (2016). R: A language and environment for mation Note, UK. statistical computing. R Foundation for Statistical Com- Uroz, S., Buee, M., Murat, C., Frey-Klett, P., and Martin, F. puting, Vienna, Austria. https://www.R-project.org/ (2010) Pyrosequencing reveals a contrasted bacterial Redford, A.J., Bowers, R.M., Knight, R., Linhart, Y., and Fierer, diversity between oak rhizosphere and surrounding soil. N. (2010) The ecology of the phyllosphere: geographic and Environ. Microbiol. Rep. 2: 281–8. phylogenetic variability in the distribution of bacteria on tree Van Dongen, W.F., White, J., Brandl, H.B., Moodley, Y., leaves. Environ. Microbiol. 12: 2885–2893. Merkling, T., Leclaire, S., et al. (2013) Age-related differ- Rossi, S., Menardi, R., and Anfodillo, T. (2006) Trephor: A ences in the cloacal microbiota of a wild bird species. New Tool for Sampling Microcores from tree stems. IAWA BMC Ecol. 13: 11. J. 27: 89–97. Vellend, M. (2010) Conceptual Synthesis in Community Salter, S.J., Cox, M.J., Turek, E.M., Calus, S.T., Cookson, Ecology. Q. Rev. Biol. 85: 183–206. W.O., Moffatt, M.F., et al. (2014) Reagent and laboratory Volkov, I., Banavar, J.R., Hubbell, S.P., and Maritan, A. contamination can critically impact sequence-based (2003) Neutral theory and relative species abundance in microbiome analyses. BMC Biol. 12: 87. ecology. Nature 424: 1035–1037. Sloan, W.T., Lunn, M., Woodcock, S., Head, I.M., Nee, S., Vorholt, J.A. (2012) Microbial life in the phyllosphere. Nat. and Curtis, T.P. (2006) Quantifying the roles of immigra- Rev. Microbiol. 10: 828–40. tion and chance in shaping prokaryote community struc- Ward, N.L., Challacombe, J.F., Janssen, P.H., Henrissat, B., ture. Environ. Microbiol. 8: 732–40. Coutinho, P.M., Wu, M., et al. (2009) Three genomes Sul, W.J., Oliver, T.A., Ducklow, H.W., Amaral-Zettler, from the phylum Acidobacteria provide insight into the L.A., and Sogin, M.L. (2013) Marine bacteria exhibit a lifestyles of these microorganisms in soils. Appl. Environ. bipolar distribution. Proc. Natl. Acad. Sci. U. S. A. 110: Microbiol. 75: 2046–56. 2342–7. Yatsunenko, T., Rey, F.E., Manary, M.J., Trehan, I., Tarpy, D.R., Mattila, H.R., and Newton, I.L.G. (2015) Character- Dominguez-Bello, M.G., Contreras, M., et al. (2012) ization of the honey bee microbiome throughout the queen- Human gut microbiome viewed across age and geogra- rearing process. Appl. Environ. Microbiol. 81: 3182–3191. phy. Nature 486: 222–7.

VC 2016 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology Reports, 8, 649–658