communications earth & environment Article
https://doi.org/10.1038/s43247-024-01895-6
Host speciesand temperaturedrivebeech
and Scots pine phyllosphere microbiota
across European forests
Check for updates
Daniela Sangiorgio 1 , Joan Cáliz 2, Stefania Mattana3,4, Anna Barceló5, Bruno De Cinti6,
David Elustondo7, Sofie Hellsten8, Federico Magnani1, Giorgio Matteucci 9, Päivi Merilä10,
Manuel Nicolas11, Dario Ravaioli1, Anne Thimonier 12, Elena Vanguelova13, Arne Verstraeten14,
Peter Waldner 12, Emilio O. Casamayor 2, Josep Peñuelas 3,4, Maurizio Mencuccini3,15 &
Rossella Guerrieri 1
Tree-microbe interactions are essential for forest ecosystem functioning. Most plant–microbe
research has focused on the rhizosphere, while composition of microbial communities in the
phyllosphere remains underexplored. Here, we use 16S rRNA gene sequencing to explore differences
between beech and Scots pine phyllospheric microbiomes at the European continental scale, map
their functional profiles, and elucidate the role of host trees, forest features, and environmental factors
such as climate and atmospheric deposition in phyllosphere microbiota assembly. We identified tree
species and the associated foliar trait (specifically carbon:nitrogen ratio) as primary drivers of the
bacterial communities. We characterized taxonomical and functional composition of epiphytic
bacteria in the phyllosphere of beech and Scots pine across an environmental gradient from
Fennoscandia to theMediterranean area, withmajor changes in temperature and nitrogen deposition.
We also showed that temperature and nitrogen deposition played a crucial role in affecting their
assembly for both tree species. This study contributes to advancing our understanding on factors
shaping phyllosphere microbial communities in beech and Scots pine at the European continental
scale, highlighting the need of broad-scale comparative studies (covering a wide range of foliar traits
and environmental conditions) to elucidate how phyllosphere microbiota mediates ecosystem
responses to global change.
Forests, covering 40million km2 of terrestrial surface1, play a crucial role as a
carbon sink, thus contributing to mitigating climate change2–4. They also
harbor a vast portion of terrestrial biodiversity1, including hiddenmicrobes
in the soil and those associatedwith trees. This diversity has led to the recent
concept of the forest microbiome, emphasizing the crucial role of micro-
organisms in influencing ecosystem functions and responses to global
change drivers5,6. However, research on the forest microbiome has
predominantly focused on the rhizosphere, leaving other habitats, such as
tree canopies (the so-called phyllosphere), underexplored.
The phyllosphere, defined as the total above-ground surfaces of
plants—including leaves, stems, flowers, and fruits— is characterized by
more dynamic and stressful conditions compared to the rhizosphere7.
Epiphytic microbes inhabiting the leaf surface face threats such as high
temperatures, heavy rainfall, drought, UV radiation exposure, desiccation,
1Department of Agricultural and Food Sciences, University of Bologna, Bologna, Italy. 2Centre of Advanced Studies of Blanes, CEAB-CSIC, Spanish Council for
Scientific Research, Blanes, Spain. 3CREAF, Cerdanyola del Vallès, Barcelona, Catalonia, Spain. 4CSIC, Global Ecology Unit CREAF-CSIC-UAB, Bellaterra,
Barcelona, Catalonia, Spain. 5Servei de Genòmica i Bioinformàtica, IBB-Parc de Recerca UAB - Mòdul B, Universitat Autònoma de Barcelona, Bellaterra,
Barcelona, Spain. 6CNR-IRET, Monterotondo Scalo, Italy. 7University of Navarra, BIOMA Institute for Biodiversity and the Environment, Pamplona, Spain. 8IVL
Swedish Environmental Research Institute, Gothenburg, Sweden. 9CNR-IBE, Roma, Italy. 10Natural Resources Institute Finland (Luke), Oulu, Finland.
11Département Recherche Développement Innovation, ONF, Office National des Forêts, Fontainebleau, France. 12WSL, Swiss Federal Institute for Forest, Snow
and Landscape Research, Birmensdorf, Switzerland. 13Centre for Forest Protection, Forest Research, Farnham, UK. 14Research Institute for Nature and Forest,
Geraardsbergen, Belgium. 15ICREA, Barcelona, Spain. e-mail: daniela.sangiorgio.93@gmail.com
Communications Earth & Environment |           (2024) 5:747 1
12
34
56
78
90
():
,;
12
34
56
78
90
():
,;
and low nutrient availability8. Additionally, these microbes are often more
sensitive to atmospheric chemistry, including pollutant concentrations
(such as nitrogen, sulfur, and organic pollutants), which they can metabo-
lize, thus contributing to the degradation of airborne pollutants8–11.
Numerous studies underlined the role of plant host species as a
selectivefilterof thephyllospheremicrobiota assembly inboth tropical12 and
temperate13,14 forests. Major epiphytic taxa identified in tree canopies in
undisturbed forests are dominated by Proteobacteria, Actinobacteria, and
Bacteroidetes12,13,15. These studies suggest that there is a coremicrobiome for
phyllospheric bacterial communities, though differences were observed
between tree species in the abundance of each of the core taxa, with iden-
tified indicator taxonomic groups associated with different host
species13,15–17. Functional profiles of phyllospheric microbes, besides allow-
ing them to survive such harsh conditions18, play a key role for their hosts by
improving nutrient and water uptake, protecting from biotic and abiotic
stresses, thus increasing plant resistance and directly acting as biological
control agents19,20.
The abiotic conditions experienced by epiphytic microbial commu-
nities living on leaf surfacesmay be expected to undergo substantial shifts at
the continental scale, with major changes in climate (temperature and
precipitation) and anthropogenic factors (atmospheric deposition). Studies
examining the latitudinal effects on forest microbial diversity and structure,
mostly focusing on the rhizosphere, have shown contrasting results. Some
studies found negative correlations with latitude21, others reported positive
correlations22, and some observed a hump-shaped trend23, with differences
in the relationships between fungi and bacteria. However, the exploration of
latitudinal effects has been less common in the case of the phyllosphere
microbiome. To the best of our knowledge, only one study has investigated
bacterial richness and composition in the phyllosphere of forests on a
continental scale in China, involving over 300 tree species from 148 genera
and 59 families24. Phyllosphere bacterial diversity and community com-
position followed a latitudinal gradient, primarily influenced by changes in
temperature andprecipitation.Our study advancesunderstandingof factors
shaping phyllospheremicrobial communities in beech and Scots pine at the
European continental scale, emphasizing the need of broad-scale com-
parative studies on foliar traits and environmental conditions to reveal how
phyllosphere microbiota mediates ecosystem responses to global change.
Atmospheric deposition of pollutants, with particular reference to
reactive nitrogen compounds, can interact with tree canopies andmicrobes
living there, thus affecting not only the availability and balance of essential
nutrients25, but also changing the acidity of thephyllospheric environment26.
Leaf associatedmicrobes contribute to both carbon27 andnitrogen cycling at
the ecosystem scale27–29, though evidence at continental scales is generally
lacking (but see ref. 30).
Pyllospheric microbial taxonomy and functional profiles are also
strongly mediated by the host species canopy and leaf traits, and therefore
they vary among tree species31. A major axis of functional variability in tree
resource economics32 is the leaf economics spectrum33, which is character-
ized by gradients in the carbon andnitrogen use, affecting leaf palatability to
biotic agents, leaf lifespan and photosynthetic return on the investment
related to leaf construction and reconstruction. For instance, high-nitrogen
acquisitive leaves (generally in deciduous species) are expected to be low-
investment and fast-return with short lifespan.Whereas conservative leaves
(such as in the case of conifers) show long life span, which comes at the cost
of very expensive high dry mass construction and low nutrient concentra-
tions. How those traits potentially affect leaf microbes’ assembly is still
poorly investigated.
Studies assessing differences in phyllosphericmicrobiota across species
have been mostly site-specific12,14,15,29 and limited to taxonomical
characterization19. Whether the ‘core microbiome’ and microbial structure
for a given species and plant functional type is maintained along an envir-
onmental gradient has not been fully explored (but see ref. 24), particularly
for European forests. In addition to the direct environmental effects, dif-
ferences in climatic factors along large environmental gradients can also
affect both foliar and canopy structure and functional traits31 for a given tree
species, thereby indirectly affecting the structure of the microbiota com-
munity assembly. Understanding whether it is the host species, the envir-
onment or rather the interaction between these two factors that drive the
microbial assembly and functional profiles in the phyllosphere at the con-
tinental scale is pivotal to elucidate global change impacts on forest health
and functioning15. Furthermore, this requires not only determining
microbiota taxonomy but also clarifying the specific functions with which
the microbes inhabiting foliar surfaces are potentially associated with.
In this study we characterized epiphytic bacterial diversity, structure,
taxonomical composition and functional profiles in the phyllosphere of
beech (Fagus sylvatica L.) and Scots pine (Pinus sylvestris L.) forests along a
large environmental gradient from Fennoscandia to the Mediterranean
area. Beech and Scots pine are two of themostwidespread and economically
important tree species in European forests34, representing two plant func-
tional types: temperate deciduous broadleaves and temperate evergreen
conifers, respectively. Along the studied gradient, temperature and nitrogen
depositionwere the two factors that varied themost, ranging from−0.58 °C
(in Finland) to 12.9 °C (in Spain), and from less than 2 kg ha⁻¹ yr⁻¹ (in
Finland) to over 15 kg ha⁻¹ yr⁻¹ (in Belgium) (Table S1). Thus, they are
expected to be very different in terms of leaf morphological and functional
traits, which should also be reflected in the different types and amounts of
nutrients available on their surfaces or those received via atmospheric
deposition30. Our specific goals were to: (i) investigate differences between
tree species foliarmicrobiomes and test whether thesewere consistent along
the gradient; (ii) map the functional profiles of microbes inhabiting the
phyllosphere; (iii) elucidate whether host species and characteristics (i.e.,
foliar carbon:nitrogen ratio, C:N, and foliar N%), forest features (i.e., alti-
tude, forest age, soil C:N, soil pH) and environmental factors (including
climate and atmospheric deposition) drive the assembly of the forest
phyllospheric microbiota. Our central hypotheses were that the microbial
diversity, structure, and functional profiles of the phyllosphere would differ
between the two host tree species, regardless of the latitudinal variation.
Additionally, we expected a significant effect of latitude (mostly via tem-
perature) and nitrogen deposition on bacterial diversity, either through a
direct impact on phyllosphere microbes or an indirect effect mediated by
foliar traits.
Results
Structure and taxonomical composition of microbial commu-
nities in beech and Scots pine
Microbial communities of the phyllospherewere clearly segregated between
tree species (R2 0.27 and p-value 0.001 PERMANOVA; Fig. 1a). Alpha
diversity estimated by the Shannon index was significantly higher in beech
than in Scots pine (Fig. 1b).
The phyllosphere prokaryotes communities in both beech and Scots
pine were largely dominated by the bacterial domain, in particular by the
classes Acidobacteriae, Actinobacteria, Alphaproteobacteria, Bacteroidia,
Deinococci, Gammaproteobacteria, and Myxococcia, all together account-
ing for at least 90% of the sequence reads in each sample (Fig. 2a). In turn,
<1% belonged to the Archaea domain, which precluded from robust sta-
tistical analysis of patterns related to Archaea. The beech leaf surfaces
showed higher relative abundance of Actinobacteria (on average 11.2 vs 2%
in beech and Scots pine, respectively) and Bacteroidia (27.6 vs 6.1%),
whereas Alphaproteobacteria (43.3 vs 54.7%) and Acidobacteria (2.5 vs
17%) showed higher relative abundance on Scots pine needle surfaces
(Fig. 2a). For Scots pine at northernmost sites in Finland (Punkaharju,
Kivalo and Sevettjarvi) Bacteroidia were not detected on the phyllo-
sphere (Fig. 2a).
The core bacterial genera, defined as those taxa detected in 100% of
samples for each species, reached 73.5% and 61.4% of the communities in
beech and Scots pine, respectively (Fig. S1). Genera belonging to the species-
specific core were more numerous in beech than in Scots pine. The number
of zOTUs specific to each tree species was high (about 50% of zOTUs
detected in each species), but rare (low abundance) within each community
(Fig. S2). In general, beech phyllospheric communities showedmore genera
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 2
with higher differential abundances, with the exception of Rhodovastum,
Granulicella, Endobacter, Bryocella, and Acidocella (Fig. 2b), which were
more abundant in Scots pine.
Differences in functional profiles of the two tree species
In order to infer microbial functional profiles, we applied the FAPROTAX-
based functional prediction, which allowed us to assign at least one func-
tional profile to around 30% of the community, on average across all com-
munities (see Methods for more details). Among the dominant
functionalities of the phyllospheric microbial community, nitrate reduction,
nitrogen fixation, methanotrophy and hydrocarbon degradation were sig-
nificantly higher in Scots pine, whereas methanol oxidation and ureolysis
were higher in beech (Fig. 3).Methylobacterium andMethylocella, in beech
and Scots pine respectively, were the genera responsible for most of the
differences observed (Fig. S3). Microorganisms with potential chemolitho-
trophic metabolisms (including methylotrophy) dominated on the foliar
surfaces, and overall showed no difference between the two tree species.
Drivers of variation in phyllosphere microbial community struc-
ture and diversity
We investigated the influence of geographic (latitude, longitude), climatic
(temperature, precipitation), host (including species and related char-
acteristics, such as foliar N% and foliar C:N as functional traits), forest
features (altitude, forest age, soil C:N, soil pH), and anthropogenic variables
Fig. 1 | Structure and diversity of phyllospheric bacterial communities in beech
and Scots pine. a NMDS ordination of phyllospheric microbial communities of
beech and Scots pine along a European gradient (n = 14 for beech and n = 22 for
Scots pine). b Boxplots representing the Shannon diversity index of the
phyllospheric microbial communities of the two tree species. Boxes represent the
median (horizontal line) and the interquartile ranges of binned values (Q25, Q75),
and the whiskers cover Q25− 1.5(Q75 –Q25) to Q75+ 1.5(Q75−Q25) with
n = 14 for beech and n = 22 for Scots pine.
Fig. 2 | Taxonomical composition of bacterial communities in beech and
Scots pine. a Taxonomic composition of the microbial communities associated to
the foliar surface of beech and Scots pine at the class level, grouping samples by the
tree species (n = 14 for beech and n = 22 for Scots pine) and forest site (n = 3, except
for Punkaharju, n = 1 and Cansiglio, n = 2), and ordered by latitude from north
(bottom) to south. b Differential abundance of zOTUs between the two tree species
shown at the genus level. zOTUs with significantly different abundances (expected
Benjamini–Hochberg corrected p-value of bothWelch’s t test <0.001 andWilcoxon
test <0.001 in Aldex2 analysis) and with a median log2 relative abundance higher
than 3 in both species were shown (n = 14 for beech and n = 22 for Scots pine).
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 3
(bulk and throughfall depositions of nitrogen and sulfur) on the phyllo-
sphericmicrobial communities.Multiple regression analysis was performed
to assess the relationships between the Shannon index and tree species and
characteristics, forest features, and environmental factors (both climatic and
anthropogenic factors). We found that the Shannon index increased with
temperature, altitude and foliar C:N, but decreased with precipitation and
soil pH (Table 1). The model explained 72% of the variance, with tem-
perature contributing themost (over 30%), followed by foliar traits and tree
species (Fig. S4). Since tree species were significant predictors in the model,
we further explored correlations between microbial Shannon diversity
indexes and all the other variables for each species independently (Table S2).
In beech, alpha diversity was negatively correlated with longitude, pre-
cipitation, and soil pH whereas positive relationships were found with
temperature, and soil C:N (Table S2). In the case of Scots pine, microbial
diversity was negatively correlated with latitude, longitude, and foliar C:N
and positively related with temperature, foliarN% andwith all the variables
belonging to the anthropogenic group (Table S2).
Mean annual temperature and longitude were the only common
variables affecting Shannon diversity in the phyllospheric microbial com-
munities of both tree species. Most of the variables included in the analyses
(geographical and climatic variables, host characteristics and species, forest
features, and anthropogenic variables) influenced the microbial assemblage
of the phyllosphere (Table S3), except SO4- S and NO₃-N TF in beech, and
soil C:N and pH in Scots pine. Variation partitioning analysis showed that
host (including species and related characteristics), forest, climatic, and
anthropogenic variables together explained 50% of the variance in phyllo-
spheric microbial community structure (Fig. 4). Host alone explained 13%
of the variance, reaching 35% when considering it simultaneously with the
other three groups. Forest and climatic factors each explained 5% and 6%of
the variance, respectively, while anthropogenic factors explained 3% when
considered alone. Altogether the four groups explained 3% of the variance.
We further explored the relationship between different variables and beta
diversity for each species by correlating variables with the main axis in the
NMDS ordination (Fig. 5). Bacterial communities in the phyllosphere dif-
fered across sites, with temperature and nitrogen deposition explaining
these differences in both species. For Scots pine, foliar C:N was also an
important factor contributing to the divergence of phyllospheric bacterial
communities across sites. Additionally, we examined the correlation
between temperature and zOTUs whose abundance was significantly dif-
ferent between beech and Scots pine according to Aldex2 analysis (Fig. 2b;
Fig. S5). In both species, the correlations were negative,meaning the relative
abundances of these zOTUs decreased at higher temperatures (Fig. S5). The
genera affected by temperature varied between species: Amnibacterium,
Kineococcus, Methylobacterium-Methylorubrum, Sphingomonas, and Spir-
osoma for beech, andBryocella, Endobacter, andGranulicella for Scots pine.
Discussion
Wehave showndistinct bacterial community structures in the phyllosphere
of beech and Scots pine, even across a broad environmental gradient
spanning boreal and temperate/Mediterraneanbiomes, thus supporting our
first hypothesis. The influence of host species identity emerged as a primary
driver of microbial structure, consistent with earlier site-specific studies in
tropical12,35 and temperate forests13,15. Indeed, leaf anatomical features, such
as absence or presence of trichomes or cuticular wax, and leaf thickness,
have been shown to contribute to shaping the composition of the phyllo-
spheric microbiota36,37.
Unlike previousfindings15where coniferous (Abies balsamea andPicea
glauca) exhibited higher alpha-diversity than deciduous species (Acer sac-
charum, Acer rubrum, Betula papyrifera), our study showed a higher
microbial diversity in beech compared to Scots pine. This discrepancy
indicates that it may be misleading to generalize results observed at the
species level to the functional type level, as microbial diversity is clearly
driven by the foliar traits of the host tree species (and hence the habitat
associated with it). Nevertheless, the distinctive bacterial structure and
taxonomical composition of the beech and Scots pine phyllosphere may be
attributed to differences in leaf and canopy structures between conifers and
deciduous species. Beech has thinner and larger leaves, while Scots pine has
thicker, smaller needles, which influences microbial exposure to environ-
mental factors and nutrient limitations38,39.
The overall taxonomic composition of the two tree species is aligned
with commonbacterial groups in thephyllosphere of various plant species20.
Regardless of the site, beech and Scots pine displayed varying percentages of
Actinobacteria, Bacteroidia, Alphaproteobacteria, and Acidobacteria.
Notably, Actinobacteria presence accounted for 2–11% of the total micro-
bial community, which is in line with previous studies in tropical, neo-
tropical and temperate forests12,15,16,40, suggesting that this is a common
taxon in the phyllosphere, regardless of the biomes and climatic regions.
Additionally, the higher presence of Actinobacteria in beech is in
Fig. 3 | Bacterial functional profiles in the phyllosphere of beech and Scots pine.
Heatmap of the functional profiles of microorganisms in the phyllosphere for beech
and Scots pine grouped by tree species (n = 14 for beech and n = 22 for Scots pine) and
forest site (n = 3, except for Punkaharju, n = 1 and Cansiglio, n = 2). Only functional
profiles with significantly different relative abundances between the two tree species
are shown (expected Benjamini–Hochberg corrected p-value ofWelch’s t test <0.05).
Table 1 | Results from the multiple regression analysis to
explore relationship between Shannon index and host (which
includes both species and host characteristics) and
environmental variables
Variable Estimate Standard
error
t- value p-value
Intercept (Beech) 3.9281 0.4228 9.290 <0.001
Species (Scots pine
- Beech)
−0.9414 0.2393 0.2393 <0.001
Temperature 2.5676 0.4381 5.861 <0.01
Precipitation −1.4497 0.5137 −2.822 <0.01
Altitude 1.7535 0.6641 2.640 <0.05
Foliar C:N 1.2869 0.5004 2.572 <0.05
Soil pH −0.7653 0.2144 −3.569 <0.01
Adjusted R2 0.7222
F-statistic 1.16 p < 0.001
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 4
accordance with the observation that deciduous trees host more of this
taxon with respect to evergreen species41.
Functionalities expressed by leaf-associatedmicroorganismshave been
proven to play a pivotal role not only in influencing plant growth, pro-
ductivity and resistance to biotic and abiotic stresses, but also in regulating
the biogeochemical cycles of carbon, nitrogen, phosphorus and
sulfur7,30,41–44.
Among the 62 functions identified by FAPROTAX, only six were
significantly different between beech and Scots pine. The low functional
variability observed among tree species might be due to the overall similar
harsh conditions that microbes experience in the phyllosphere14. Never-
theless, interesting differences emerged, with functions related to ureolysis
andmethanol oxidation, more represented in beech than in Scots pine. The
phyllosphere is a particularly privileged habitat of methanol-utilizing bac-
teria, such as Methylobacterium, which play a key role in the methanol
emission mitigation thanks to their direct consumption45. Methanol is
mostly released as a by-product of demethylesterification of homo-
galacturans of cell wall pectins and is associated with cell wall maturation46.
Indeed, young growing leaves have been found to emit higher amounts of
methanol relative tomature ones47, whichmight be the reason for the higher
abundanceofmicrobial functions related tomethanol oxidationobserved in
beech17. In addition, Scots pine phyllosphere were richer inmicrobes able to
perform methanotrophy, nitrogen-fixation, nitrate reduction, and hydro-
carbon degradation. In some methanotrophs, nifH gene encoding for N2
fixation can be present48,49, which partially explain the co-occurrence of the
two functions (methanotrophy and N2 fixation). Methanotrophy is a
microbial function found in several environments (wetlands, marshes, rice
paddies, landfills, aquatic systems) and is essential for reducing the amount
of methane emitted to the atmosphere. The phyllosphere harbors a lower
abundance of methanotrophs compared to the soil, probably due to the
intrinsic lownutrient availabilityon foliar surfaces for bacterial sustenance50.
We can hypothesize that the reduced presence of methanotrophs and
N-fixing bacteria on beech vs. Scots pine could be related to the lowerC:N in
beech leaves. Indeed, N-fixation can be favored under C-rich but N-poor
substrates, thus at high foliar C:N51,52, whereas free-living N fixation has
shown to be suppressed under high N conditions53.
We also found Scots pine needle surfaces to be richer in microbial
functionalities linked to hydrocarbon degradation. It has been found that
the high concentration of atmospheric pollutants deposited on leaves,
including PM10 and polycyclic aromatic hydrocarbons (PAHs), favor the
selection of hydrocarbondegradingbacteria inhabiting thephyllosphere54,55.
Additionally, Scots pine showing higher functionality related to hydro-
carbon degradation could be linked to the foliarmorphology, i.e., high foliar
thickness, presence of waxes and longer lifespan compared to beech, which
altogethermakes needlesmore effective in capturing pollutants56 in general.
It is important to emphasize that in silico functional predictions should
be evaluated with caution. Tools like FAPROTAX, Picrust2, Tax4Fun,
BugBase, and others57 still have limited representation of the entire micro-
bial community obtained by ribosomal genes analyses. Transcriptomic
analysis of microbial communities remains the most powerful and reliable
method for elucidating the functional profiles expressed within these
ecosystems58,59. This approach offers unparalleled insights into the dynamic
processes and gene expression profiles that drive microbial functions,
providing a comprehensive understanding of the roles these communities
play in situ.
Plant microbiomes are influenced by various factors, including host
phenotype, species, genotype, environment, and microbe-microbe
interactions60,61. Understanding the impact of biotic and abiotic drivers on
the plant holobiont is crucial for predicting the effects of climate change on
forest microbes and the ecological services they provide. In this study, we
sought to determine the extent to which environmental variables influence
the structure of the phyllospheric bacterial microbiota of two major Eur-
opean species, beech and Scots pine, at the continental scale. Along the
investigated gradient, there are substantial changes in temperature (from
Fig. 4 | Drivers of variation in beech and Scots pine phyllosphere bacterial
community structure. Variation Partitioning Analysis (VPA) of microbial com-
munity structure with explained variation by host (species and foliar C:N), forest
features (altitude, forest age and soil C:N), climatic (temperature and precipitation),
and anthropogenic variables (TN BD).
Fig. 5 | Structure of phyllospheric bacterial communities arranged per site,
highlighting key variables. NMDS ordination of phyllospheric microbial com-
munities for beech (n = 14) (a) and Scots pine (n = 22) (b) grouped by study sites
(n = 3, except for Punkaharju, n = 1 and Cansiglio, n = 2). The arrows above indicate
the Spearman’s correlation coefficient ( > 0.7) between the ordination scores of the
first axis and variables.
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 5
−0.58 to 12.9 °C), as well as atmospheric deposition. In this latter case, we
observed the highest values inCentral Europe (Belgium, Switzerland), while
the lowest in the Fennoscandian countries. This is also reflected in soil C:N
and the foliar C:N stoichiometry, in particular for the Scots pine, i.e., lower
C:N ratio at high N deposition sites.
The combination of host factors (tree species and foliar C:N as a
functional trait), forest attributes (forest age, altitude, soil C:N), climatic
variables (temperature and precipitation), and anthropogenic factors
(deposition of total N) explained 50% of the observed variation in phyllo-
spheric microbial assemblages. The relatively low variance explained by
these variables is consistent with a study assessing diversity at a continental
scale in China24, suggesting that other unexplored biotic and abiotic factors,
such as specific leaf features (e.g., nutrient concentrations, secondary
metabolite emissions, leaf morphological traits, or intercepted light), could
likely play a role in affecting beta diversity62.
Although we cannot rule out a latitudinal effect related to inter-
continental atmospheric bacterial deposition63, temperature emerged as a
common environmental driver influencing both alpha and beta diversity
indexes. Specifically, alpha diversity increased with temperature (and
decreased with latitude or altitude) along the investigated gradient in Eur-
ope. This result aligns with other studies assessing phyllospheric microbial
diversity along altitudinal gradients64, though it contrastswith a continental-
scale study in China that reported a hump-shaped biodiversity pattern in
relation to latitude24. Nevertheless, our findings support thewell-established
role of temperature in affecting microbial physiology and metabolism65.
Temperature also appears to influence the relative abundance of taxa spe-
cific to beech and Scots pine along the gradient, though further exploration
is needed to determine their ecological roles in the phyllosphere and their
links to the host species.
The effects of nitrogen deposition on forest growth, biogeochemical
processes66–68, and vegetation diversity69–71 have been widely explored in the
literature. Recent studies have shown that increased nitrogen deposition
plays a crucial role in shaping ectomycorrhizal structure and functionality in
the rhizosphere72,73, though its effect on phyllospheric microbial commu-
nities has been less explored. In our study, besides temperature, nitrogen
deposition was identified as a key driver of beta diversity in both beech and
Scots pine, suggesting that bacterial assemblages are sensitive not only to
climatic changes but also to atmospheric chemistry conditions. For Scots
pine in particular, we observed a positive correlation between microbial
alpha diversity and foliarN%, aswell as a key role of foliar C:N in explaining
beta diversity. These results, together with the higher presence of potential
bacterial functionality associated with nitrogen-fixation and nitrogen-
reduction in Scots pine, suggests that nitrogen-limited conditions in the case
of Scots pine needles vs. beech leaves may trigger the presence of microbes
able to use atmospheric nitrogen for their metabolism10,74.
In conclusion, we identified host species identity and the associated
foliar trait (C and N stoichiometry) as the primary driver of the phyllo-
spheric microbiota, both in terms of taxonomy and functional profiles.
Moreover, we showed that temperature and nitrogen deposition played a
pivotal role in explaining assembly of phyllosphere bacterial communities
for both tree species along the large latitudinal gradient in Europe. Our
results highlight the complex, yet synergic – interactions among biotic and
environmental drivers in affecting forest phyllopsheric microbiota and
functional profiles at the European continental scale. Extending the meta-
genomic approaches to other tree species, covering a wider range of foliar
traits and assessing their functional profiles are important steps forward to
elucidate causes of variations in phyllospheric microbial communities and
their role in nutrient cycling, stress resistance, and other processes under-
pinning ecosystem functioning under global change.
Materials and methods
Sites description and foliar samples collection
Thirteen forested sites within the Level II ICP Forests network (http://icp-
forests.net/), including as dominant species (Table S4) two of the most
common European tree species (Fagus sylvatica L. and Pinus sylvestris L.),
were selected to span a wide range of environmental conditions, including
forest features, climate and atmospheric nitrogen and sulfur depositions
(Table S1).
At each site, professional tree climbers collected foliar samples from
five trees chosen amongst those already considered for nutrient analysis in
the ICP Forests network. Three shoots from each tree were sampled in the
upper, middle, and lower third of the canopy in August 2016, except for
Sweden and Finland where the samples were collected in October 2016 and
August 2017, respectively. To avoid the contact between foliage and the
ground (and possible contamination with soil microbes), shoots were
sampled from the canopy and were immediately placed in labeled sterilized
bags, which were sealed when the tree climber was still in the canopy. The
sealed bags were then dropped to the forest floor and immediately placed in
a box containing dry ice. The foliar samples were stored in the laboratory at
−20 °C until the microbial DNA extraction.
Forest and environmental data
For each site, we considered the following variables: Forest features (altitude,
forest age, soil carbon:nitrogen ratio (soil C:N) and soil pH) and host
characteristics (foliar nitrogen concentration in percent of drymass (foliarN
%) and foliar carbon:nitrogen ratio (foliar C:N)). Soil C:N was obtained
fromthe top10 cm,while foliar samples collected in2016 atmost of the sites,
except for the Finnish sites where sampling was carried out in 2017. Pro-
tocols for soil sampling and laboratory analyses at ICP Forests sites are
described in ref. 75 and ref. 76, respectively. Site information included
several variables, ranging from longitude and latitude (named as geo-
graphic) and environmental variables, including climate and anthropogenic
factors. Within the climate factors, we included mean annual temperature
and total annual precipitation for the year samplingwas carriedout,whereas
anthropogenic factors included bulk (BD) and throughfall (TF) depositions
of inorganic nitrogen (NH4-N, NO3-N, total N (TN)) and sulfur (SO4- S).
TNwas obtained as the sum of inorganic N and dissolved organic nitrogen.
For more details on atmospheric deposition fluxes quantification, please
refer to ref. 77 and ref. 30.
DNA extraction and sample preparation for metabarcoding
analysis
Of the five trees originally considered for sampling shoot, only three were
used for the analysis of phyllospheric microbiota, and the remaining two
were used for repeating DNA extraction when not enough microbial DNA
(e.g., the forest sites in Sweden and Finland) was available. Microbial DNA
extraction and sample preparation were performed as described in ref. 30.
Briefly, epiphyticmicrobialDNAwasobtained from5–6 g (for beech leaves)
and 8–10 g (for Scots pine needles) of foliage randomly collected from each
of the three shoots sampled per tree and placed (as a composite sample for
each tree) in sterile 50-mL Falcon tubes20. This allowed to have a canopy-
level information on epiphytic bacteria. Thirty-fivemilliliters of 1:50 diluted
Redford buffer wash solution (1M Tris·HCl, 0.5M Na EDTA, and 1.2%
CTAB12)was added to each tube,whichwas then stirred for 5min.Then, the
washing solution containing the leaf epiphyteswas centrifuged for 30min at
3000 g. Theobtainedpelletwas transferred to 2‐mlMOBIOPowerSoil bead
beating tubes for DNA extraction, which was conducted following the
manufacturer’s instructions (DNeasy PowerSoil Kit, Qiagen, Benelux BV;
previously the PowerSoil DNA isolation kit fromMo Bio laboratories). The
16S rRNA V5-V6 region was targeted by 799 F and 1115 R, cyanobacteria
and chloroplast excluding primers12. Overhang adapters (forward 5′TC
GTCGGCAGCGTCAGATGTGTATAAGAGACAGAACMGGATTA-
GATACCCKG and reverse 5′GTCTCGTGGGCTCGGAGATGTGTA-
TAAGAGACAGAGGGTTGCGCTCGTTG)were attached to the primers,
as recommended in the Illuminamanual for 16SMetagenomic Sequencing
Library Preparation. Reaction for the first amplification had a total volume
of 25 μL consisting of 1–2 μL of Amplicon PCR Forward Primer (5 μM),
1–2 μL of Amplicon PCRReverse Primer (5 μM), 2.5 μL ofmicrobial DNA,
12.5 μL of 2× KAPA HiFi HotStart ReadyMix (Kapa Biosystems), and
MilliQ water. Reactions were carried out following the Illumina procedure:
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 6
95 °C for 3min and then 34 cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C
for 30 s, with a final extension at 72 °C for 30min. LabChip electrophoresis
indicated that the amplicons were 400 bp long. The reactions were purified
using CleanPCR beads (CleanNa) and ethanol to remove free primers and/
or primer dimers. A second PCR was performed with the cleaned PCR
product as template to attachdual indices and Illumina sequencing adapters
using the Nextera XT Index primers and following Illumina protocols.
Multiplexed 16S libraries were prepared bymixing 5 μL of each reaction at a
concentration of 8 pM with 30% PhiX (the internal DNA control from
Illumina). The microbial DNA was quantified using a NanoDrop spectro-
photometer (Thermo Fisher Scientific, Wilmington, DE, USA). Aliquots of
microbial DNAobtained as described above were used to prepare amplicon
libraries for Illumina 16S rRNA gene sequencing using an Illumina MiSeq
instrument with a 500-cycle cartridge. DNA extraction, amplicon pre-
paration, and sequence analysis were carried out at Servei de Genòmica i
Bioinformàtica (Universitat Autonoma de Barcelona, Spain).
DNA sequencing processing
Raw 16S rRNA gene sequences were processed by following the UPARSE
pipeline78 (Edgar, 2013) as follows (i) forward and reverse reads were paired
to obtain consensus sequences, (ii) low‐quality sequences were discarded
based on the expected error filtering, set to 0.5 and (iii) sequences were
denoised (error-correction) anddefined intooperational taxonomic units at
100% identity, that is zero-radius OTUs (zOTUs)79. SINA80 and SILVA 138
database81 were used for the taxonomic assignment. Chloroplast, mito-
chondrial and unclassified sequences were excluded from further analyses,
and a total of 2’827’137highquality readswere used. FunctionalAnnotation
of Prokaryotic Taxa (FAPROTAX) tool was used to determine the potential
functional profile of each zOTU82. The functional profiles of microorgan-
isms identified in the phyllosphere of beech and Scots pinewithhigh relative
abundanceswere all related to commonmetabolic functions in both species.
For diversity and community structure analyses, the original zOTU table
was rarefied and set to a minimum depth of 4500 reads per sample in order
to minimize biases for differences in sampling effort. One replicate of the
forest site Cansiglio was removed due to the low sequencing depth. All the
three samples of the site located in Collelongo were removed from the
dataset because late frost in 2016 affected the beech forest in Collelongo,
causing complete defoliation, which was followed by a second green-up of
the canopy in June83. Analysis of themicrobial community structure clearly
showed a separation of microbial communities from the rest of the
sites (Fig. S6).
Statistical analysis
Shannon diversity was calculated using the ‘diversity’ function from the
vegan package84. To identify factors affecting Shannon diversity, we per-
formed multiple linear regression analysis, using forest features (altitude,
forest age, soil C:N, and soil pH), environmental (including climatic and
anthropogenic factors), and host variables (including tree species and
related characteristics) as predictors. We applied a two-directional stepwise
model selection based on the Akaike information criterion (AIC) and
ensured that the variance inflation factor (VIF) for all variables remained
below 10. The assumptions of the model were checked using the sjPlot
package85.
The relative importance of predictors in the final model was assessed
using the Lindeman,Merenda, andGold (lmg)method, which partitions R²
by averaging over all orders, through the ‘calc.relimp’ function in the
relaimpo package86. Spearman correlations between the Shannon diversity
index and environmental variables were computed separately for the beech
and Scots pine datasets.
Continuous variables considered for the forest and environmental data
were first rescaled using the formula (x - xmin) / (xmax - xmin), where x
represents the given variable. To reduce collinearity inmultivariate analyses,
we calculated pairwise Pearson correlation coefficients between variables
and removed thosewith a correlation coefficient greater than0.9.As a result,
foliarN%, all anthropogenic variables inTF, aswell asNH₄-NandNO₃-N in
BD, were excluded from the analysis. Geographic coordinates were also not
considered due to their relationship with climatic factors.
A hypothesis contrast test (Student’s t-test) was used to assess the effect
of tree species on Shannon diversity.
For community composition, we calculated a Bray–Curtis distance
matrix based on Hellinger transformed community data87. Nonmetric
multidimensional scaling (NMDS) analysis was performed to show the
variation in composition between the two tree species, and across the study
sites for each species, where NMDS scores of the first axis were correlated
with continuous variables of forest and environmental data by calculating
the Spearman’s correlation coefficient. The relation of forest, environmental
and host variables on the microbial structure was analyzed separately by
permutational analysis of variance (PERMANOVA). Each variable was
analyzed separately, with community data transformed by Hellinger and
999 permutations, as implemented in the ‘adonis’ function in vegan pack-
age.We also tested towhich extent the variation in the community structure
was related to four groups of variables: (i) host (species, foliar N%, foliar
C:N), (ii) forest, (altitude, forest age, soil C:N, soil pH) (iii) climatic (tem-
perature and precipitation), and (iv) anthropogenic variables (bulk (BD)
and throughfall (TF) depositions of inorganic nitrogen (NH4-N, NO3-
N,total N (TN)) and sulfur (SO4- S)). For this purpose, we carried out a
variation partitioning analysis (VPA) of the Bray–Curtis dissimilarities
matrix as a function of these four groups by the ‘varpart’ function in
vegan. For each group, variables were selected by two directional permu-
tation tests using the ‘ordistep’ function to avoid multicollinearity in the
VPA model.
The identification of zOTUs and functional profiles with significantly
different abundances between beech and Scots pine was carried out using
‘aldex.clr’ function for compositional data, in Aldex2 package88. We calcu-
lated the Spearman’s correlation between temperature and relative abun-
dance of differentially abundant zOTUs for beech and Scots pine,
respectively. The false discovery rate (fdr) was applied to adjust probability
(p) values for multiple comparisons.
Plots and statistical analyses were performed in R studio89 by using
ggplot290, ggpubr91, tidyverse92, and ggfortify93 packages.
Reporting summary
Further information on research design is available in the Nature Portfolio
Reporting Summary linked to this article.
Data availability
Genetic data from 16S sequence analyses are available in the Sequence
Reading Archive at the National Center for Biotechnology Information
under accession no. PRJNA859654. All variables included in the statistical
analyses are reported in Table S1.
Code availability
The coding involved in this study is for statistical analyses, using the specific
packages described in the ‘Statistical analyses’.
Received: 10 March 2024; Accepted: 11 November 2024;
References
1. FAO. 2022. In Brief to The State of the World’s Forests 2022. Forest
pathways for green recovery and building inclusive, resilient and
sustainable economies. Rome, FAO. https://doi.org/10.4060/
cb9363en.
2. Canadell, J. G. & Raupach,M. R.Managing forests for climate change
mitigation. Science 320, 1456–1457 (2008).
3. Grassi, G. et al. The key role of forests in meeting climate targets
requires science for crediblemitigation.Nat. Clim.Change7, 220–226
(2017).
4. Friedlingstein, P. et al. Global carbon budget 2023. Earth Syst. Sci.
Data 15, 5301–5369 (2023).
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 7
5. Baldrian, P. Forest microbiome: diversity, complexity and dynamics.
FEMS Microbiol. Rev. 41, 109–130 (2017).
6. Baldrian, P. et al. Forest microbiome and global change. Nat. Rev.
Microbiol. 21, 1–15 (2023).
7. Lindow,S. E. &Leveau, J.H. J. Phyllospheremicrobiology.Curr.Opin.
Biotechnol. https://doi.org/10.1016/S0958-1669(02)00313-0 (2002).
8. Lindow, S. E. & Brandl, M. T. Microbiology of the phyllosphere. Appl.
Environ. Microbiol. 69, 1875–1883 (2003).
9. Inácio, J. et al. Estimation and diversity of phylloplane mycobiota on
selected plants in a Mediterranean–type ecosystem in Portugal.
Microb. Ecol. 44, 344–353 (2002).
10. Wang, Z. et al. Response of bacterial communities and plant-
mediated soil processes to nitrogen deposition and precipitation in a
desert steppe. Plant Soil 448, 277–297 (2020).
11. Yao, H. et al. Phyllosphere epiphytic and endophytic fungal
community and network structures differ in a tropical mangrove
ecosystem.Microbiome 7, 1–15 (2019).
12. Kembel, S. W. et al. Relationships between phyllosphere bacterial
communities and plant functional traits in a neotropical forest. Proc.
Natl. Acad. Sci. USA 111, 13715–13720 (2014).
13. Redford, A. J. et al. The ecology of the phyllosphere: geographic and
phylogenetic variability in the distribution of bacteria on tree leaves.
Environ. Microbiol. 12, 2885–2893 (2010).
14. Lajoie,G. et al. Adaptivematchingbetweenphyllospherebacteria and
their tree hosts in a neotropical forest.Microbiome 8. https://doi.org/
10.1186/s40168-020-00844-7 (2020).
15. Laforest-Lapointe, I. et al. Host species identity, site and time drive
temperate tree phyllosphere bacterial community structure.
Microbiome 4. https://doi.org/10.1186/s40168-016-0174-1 (2016).
16. Kim, M. et al. Distinctive phyllosphere bacterial communities in
tropical trees.Microb. Ecol. 63, 674–681 (2012).
17. Lambais, M. R. et al. Phyllosphere metaproteomes of trees from the
Brazilian atlantic forest show high levels of functional redundancy.
Micro Ecol. 73, 123–134 (2017).
18. Mercier, J. &Lindow,S.E.Roleof leaf surfacesugars in colonizationof
plants by bacterial epiphytes. Appl. Environ. Microbiol. 66, 369–374
(2000).
19. Lemanceau, P. et al. Let the core microbiota be functional. Trends
Plant Sci. 22, 583–595 (2017).
20. Bashir, I. et al. Phyllosphere microbiome: diversity and functions.
Microbiol. Res. 254, 126888 (2022).
21. Fu, X. et al. Latitude variations of soil bacterial community diversity
and composition in three typical forests of temperate, northeastern of
China. Front. Earth Sci. 10, 1096931 (2023).
22. Tian, J. et al. Soil organic matter availability and climate drive
latitudinal patterns in bacterial diversity from tropical to cold
temperate forests. Funct. Ecol. 32, 61–70, https://besjournals.
onlinelibrary.wiley.com/doi/full/10.1111/1365-2435.12952 (2018).
23. Liu, S. et al. Decoupled diversity patterns in bacteria and fungi across
continental forest ecosystems.Soil Biol. Biochem. 144, 107763 (2020).
24. Wang, Z. et al. Diversity and biogeography of plant phyllosphere
bacteria are governed by latitude-dependent mechanisms.N. Phytol.
240, 1534–1547 (2023).
25. Guerrieri, R. et al. Canopy exchange and modification of nitrogen
fluxes in forest ecosystems. Curr. Forestry Rep. 7, 115–137 (2021).
26. Augustaitis, A. et al. The seasonal variability of air pollution effects on
pine conditions under changing climates. Eur. J. For. Res. 129,
431–441 (2010).
27. Laforest-Lapointe, I. et al. Leaf bacterial diversity mediates plant
diversity and ecosystem function relationships. Nature 546, 145–147
(2017).
28. Fürnkranz, M. et al. Nitrogen fixation by phyllosphere bacteria
associated with higher plants and their colonizing epiphytes of a
tropical lowland rainforest of Costa Rica. ISME J. 2, 561–570 (2008).
29. Guerrieri, R. et al. Partitioning between atmospheric deposition and
canopy microbial nitrification into throughfall nitrate fluxes in a
Mediterranean forest. J. Ecol. 108, 626–640 (2020).
30. Guerrieri, R. et al. Substantial contribution of tree canopy nitrifiers to
nitrogen fluxes in European forests. Nat. Geosci. 17, 130–136 (2024).
31. Pangga, I. et al. Climate change impactsonplant canopyarchitecture:
implications for pest and pathogenmanagement. Eur. J. Plant Pathol.
135, 595–610 (2013).
32. Reich, P. B. The world-wide ‘fast–slow’ plant economics spectrum: a
traits manifesto. J. Ecol. 102, 275–301 (2014).
33. Wright, I. J. et al. Theworldwide leaf economics spectrum.Nature428,
821–827 (2004).
34. San-Miguel-Ayanz, J. et al. European Commission, Joint Research
Centre. European atlas of forest tree species (2022).
35. Li, M. et al. Phyllosphere bacterial and fungal communities vary with
host species identity, plant traits and seasonality in a subtropical
forest. Environ. Microbiome 17, 1–13 (2022).
36. Reisberg, E. E. et al. Phyllosphere bacterial communities of trichome-
bearing and trichomeless Arabidopsis thaliana leaves. Antonie Van.
Leeuwenhoek 101, 551–560 (2012).
37. Reisberg, E. E. et al. Distinct phyllosphere bacterial communities on
Arabidopsis wax mutant leaves. PLoS One 8, e78613 (2013).
38. Leveau, J. H. A brief from the leaf: latest research to inform our
understanding of the phyllosphere microbiome.Curr. Opin. Microbiol
49, 41–49 (2019).
39. Duan, Y. et al. Forest top canopybacterial communities are influenced
by elevation and host tree traits. Environ. Microbiome 19, 21 https://
www.ncbi.nlm.nih.gov/pmc/articles/PMC10998314/ (2024).
40. Imperato, V. et al. Characterisation of the Carpinus betulus L.
Phyllomicrobiome in urban and forest areas. Front. Microbiol. 10,
1110 (2019).
41. Uroz, S. et al. Ecology of the forestmicrobiome: highlights of temperate
and boreal ecosystems. Soil Biol. Biochem. 103, 471–488 (2016).
42. Laforest-Lapointe, I. & Whitaker, B. K. Decrypting the phyllosphere
microbiota: progress and challenges. Am. J. Bot. 106, 171–173
(2019).
43. Xiang, Q. et al. Microbial functional traits in phyllosphere are more
sensitive to anthropogenic disturbance than in soil. Environ. Pollut.
265, 114954 (2020).
44. Vorholt, J. A. Microbial life in the phyllosphere. Nat. Microbiol. 10,
828–840 (2012).
45. Kanukollu, S. et al. Methanol utilizers of the rhizosphere and
phyllosphere of a common grass and forb host species. Environ.
Microbiomes 17, 1–15 (2022).
46. Delmotte, N. et al. Community proteogenomics reveals insights into
the physiology of phyllosphere bacteria. Proc. Natl Acad. Sci. USA
106, 16428–16433 (2009).
47. Nemecek-Marshall, M. et al. Methanol emission from leaves
(enzymatic detection of gas-phasemethanol and relation ofmethanol
fluxes to stomatal conductance and leaf development). Plant Physiol.
108, 1359–1368 (1995).
48. Larmola, T. et al. Methanotrophy induces nitrogen fixation during
peatland development. Proc. Natl. Acad. Sci. USA 111, 734–739
(2014).
49. Zhu, Y.-G. et al. Harnessing biological nitrogen fixation in plant leaves.
Trends Plant Sci. 28, 1391–1405 (2023).
50. Iguchi, H. et al. Distribution of methanotrophs in the phyllosphere.
Biosci. Biotechnol. Biochem. 76, 1580–1583 (2012).
51. Zheng, M. Stoichiometry controls asymbiotic nitrogen fixation and its
response to nitrogen inputs in a nitrogen-saturated forest.Ecology99,
2037–2046 (2018).
52. Salemaa, M. et al. N2 fixation associated with the bryophyte layer is
suppressed by low levels of nitrogen deposition in boreal forests. Sci.
Total Environ. 653, 995–1004 (2019).
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 8
53. Dynarski, K. A. & Houlton, B. Z. Nutrient limitation of terrestrial free-
living nitrogen fixation. N. Phytol. 217, 1050–1061 (2018).
54. Franzetti, A. et al. Plant-microorganisms interaction promotes
removal of air pollutants in Milan (Italy) urban area. J. Hazard. Mater.
384, 121021 (2020).
55. Gandolfi, I. et al. Diversity and hydrocarbon-degrading potential of
epiphytic microbial communities on Platanus x acerifolia leaves in an
urban area. Environ. Poll. 220, 650–658 (2017).
56. Chen, L. et al. Variation in tree species ability to capture and retain
airborne fine particulate matter (PM2.5). Sci. Rep. 7, 3206 (2017).
57. Djemiel, C. et al. Inferringmicrobiota functions from taxonomic genes:
a review. GigaScience 11, giab090 (2022).
58. Gomez-Silvan, C. et al. A comparison of methods used to unveil the
genetic and metabolic pool in the built environment.Microbiome 6,
1–16 (2018).
59. Hempel, C. A. et al. Metagenomics versus total RNA sequencing:
most accurate data-processing tools, microbial identification
accuracyandperspectives for ecological assessments.NucleicAcids
Res. 50, 9279–9293 (2022).
60. Chaudhry,V.etal. Shaping the leafmicrobiota: plant–microbe–microbe
interactions. J. Exp. Bot. 72, 36–56 (2021).
61. Sangiorgio, D. et al. Taxonomical and functional composition of
strawberry microbiome is genotype-dependent. J. Adv. Res. 42,
189–204 (2022).
62. Hamonts, K. et al. Field study reveals core plant microbiota and
relative importance of their drivers. Environ. Microbiol. 20, 124–140
(2018).
63. Casamayor, E. O., Cáliz, J., Triadó-Margarit, X. & Pointing, S. B.
Understanding atmospheric intercontinental dispersal of harmful
microorganisms. Curr. Opin. Biotechnol. 81, 102945 (2023).
64. Wang, X. et al. Leaf traits and temperature shape the elevational
patterns of phyllosphere microbiome. J. Biogeogr. 50, 2135–2147
(2023).
65. Knapp, B. D. & Huang, K. C. The effects of temperature on cellular
physiology. Ann. Rev. Biophys. 51, 499–526 (2022).
66. Gundersen, P. et al. Impact of nitrogen deposition on nitrogen cycling
in forests: a synthesis of NITREX data. For. Ecol. Manag. 101, 37–55
(1998).
67. Etzold, S. et al. Nitrogen deposition is the most important
environmental driver of growth of pure, even-aged and managed
European forests. For. Ecol. Manag. 458, 117762 (2020).
68. Guerrieri, R. et al. Climate and atmospheric deposition effects on
forestwater-useefficiencyandnitrogenavailability acrossBritain.Sci.
Rep. 10, 12418 (2020).
69. Phoenix, G. K. et al. Atmospheric nitrogen deposition in world
biodiversity hotspots: the need for a greater global perspective in
assessing N deposition impacts. Glob. Change Biol. 12, 470–476
(2006).
70. Bobbink, R. &Hicks,W. Factors affectingnitrogendeposition impacts
on biodiversity: an overview. In Nitrogen Deposition, Critical Loads
and Biodiversity (eds Sutton, M., Mason, K., Sheppard, L., Sverdrup,
H., Haeuber, R. & Hicks, W.) https://doi.org/10.1007/978-94-007-
7939-6_14 (Springer, 2014).
71. Payne, R. J. et al. Nitrogen deposition and plant biodiversity: past,
present, and future. Front. Ecol. Environ. 15, 431–436 (2017).
72. vander Linde, S. et al. Environment andhost as large-scale controls of
ectomycorrhizal fungi. Nature 558, 243–248 (2018).
73. Moore, J. A. et al. Fungal community structure and function shiftswith
atmospheric nitrogen deposition. Glob. Change Biol. 27, 1349–1364
(2021).
74. Sardans, J. et al. Foliar elemental composition of European forest tree
speciesassociatedwith evolutionary traits andpresent environmental
and competitive conditions. Glob. Ecol. Biogeogr. 24, 240–255
(2015).
75. Cools, N. & De Vos, B. Part X.: Sampling and Analysis of Soil. Version
2020-1. In:UNECE ICPForestsProgrammeCo-ordinatingCentre (ed.):
Manual onmethods and criteria for harmonized sampling, assessment,
monitoring and analysis of the effects of air pollution on forests. Thünen
Institute of Forest Ecosystems, Eberswalde, Germany, 29 p. + Annex
http://www.icp-forests.org/manual.htm (2020).
76. Rautio, P. et al. Part XII: Sampling and Analysis of Needles and
Leaves. Version 2020-3. In: UNECE ICP Forests Programme Co-
ordinating Centre (ed.): Manual on methods and criteria for
harmonized sampling, assessment, monitoring and analysis of the
effects of air pollution on forests. Thünen Institute of Forest
Ecosystems, Eberswalde, Germany, 16 p. + Annex http://www.icp-
forests.org/Manual.htm (2020).
77. Clarke, N. et al. Part XIV: Sampling and Analysis of Deposition. Version
2022-1. In:UNECE ICPForestsProgrammeCo-ordinatingCentre (ed.):
Manual onmethods and criteria for harmonized sampling, assessment,
monitoring and analysis of the effects of air pollution on forests. Thünen
Institute of Forest Ecosystems, Eberswalde, Germany, 34 p. + Annex
http://www.icp-forests.org/Manual.htm (2022).
78. Edgar, R. C. UPARSE: highly accurateOTUsequences frommicrobial
amplicon reads. Nat. Methods 10, 996–998 (2013).
79. Edgar, R. C. UNOISE2: improved error-correction for Illumina 16Sand
ITS amplicon sequencing. Preprint at bioRxiv. https://doi.org/10.
1101/081257 (2016).
80. Pruesse, E. et al. SINA: accurate high-throughput multiple sequence
alignment of ribosomal RNA genes. Bioinformatics 28, 1823–1829
(2012).
81. Quast, C. et al. The SILVA ribosomal RNA gene database project:
improved data processing and web-based tools. Nucleic Acids Res.
41, D590–D596 (2013).
82. Louca, S. et al. Decoupling function and taxonomy in the global ocean
microbiome. Science 353, 1272–1277 (2016).
83. D’Andrea, E. et al. Winter’s bite: beech trees survive complete
defoliation due to spring late-frost damage by mobilizing old C
reserves. N. Phytol. 224, 625–631 (2019).
84. Oksanen, J. et al. 2009. The vegan Package. Community ecology
package, version 2.9 1–295 (2013).
85. Lüdecke, D. sjPlot: data visualization for statistics in social science.R.
package version 2, 1–106 (2021).
86. Grömping, U. Relative importance for linear regression in R: the
package relaimpo. J. Stat. Softw. 17, 1–27 (2007).
87. Legendre, P. & Gallagher, E. D. Ecologically meaningful transformations
for ordination of species data. Oecologia 129, 271–280 (2001).
88. Fernandes, A. D. et al. ANOVA-like differential expression (ALDEx)
analysis for mixed population RNA-Seq. PloS one 8, e67019 (2013).
89. R. Core Team, 2021. R: A language and environment for statistical
computing. Published online 2020.
90. Wickham,H. Package tidyverse. Easily Install andLoad the ‘Tidyverse
(2017).
91. Kassambara, A. Package ‘ggpubr.’ R package version 0.1 6 (2020).
92. Wickham, H. ggplot2.WIREs Comput. Stat. 3, 180–185 (2011).
93. Tang, Y., Horikoshi, M. & Li, W. ggfortify: unified interface to visualize
statistical results of popular R packages. R. J. 8, 474 (2016).
Acknowledgements
Metagenomic analyses presented in this study were supported by EU
funding from the MSCA individual fellowship (NITRIPHYLL no. 705432
awarded to R.G.; J.C. and E.O.C. were supported by project AEROSMIC
PID2021- 127701NB-I00 from theSpanish Agency of Research (AEI-MICIN,
Spain) andEuropean funding (ERDF)EuropeanRegionalDevelopmentFund
to E.O.C.; This study greatly benefited from the large efforts of the site PIs
and collaborators coordinating long-termmonitoring within the ICP Forests
network.We thank the two anonymous reviewers for their comments on the
earlier version of the manuscript.
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 9
Author contributions
R.G., M.M., and J.P. conceived the study. R.G. andM.M. led the experimental
design. A.V., E.V., P.W., A.T., D.E., B.D.C., S.H., P.M., M.N., F.M., D.R., and
G.M. led foliar sampling. R.G. was responsible for extracting microbial DNA,
with the supervision of A.B. and S.M.; A.B. sequenced the DNA using an
Illumina platform. J.C. processed the raw DNA sequences obtained from the
Illumina platform. D.S. conducted bioinformatic analyses and finalized figures
with the support of J.C. andE.O.C.; D.S. andR.G. prepared theoriginal draft of
the manuscript, with input from J.C., M.M., E.O.C., and J.P.; All authors
discussed the results and commented on the manuscript.
Competing interests
RossellaGuerrieri is anEditorial BoardMember forCommunicationsEarth &
Environment, but was not involved in the editorial reviewof, nor the decision
topublish this article. The remaining authorsdeclarenocompeting interests.
Additional information
Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s43247-024-01895-6.
Correspondence and requests for materials should be addressed to
Daniela Sangiorgio.
Peer review information Communications Earth & Environment thanks
Birgit Wassermann and the other, anonymous, reviewer(s) for their
contribution to the peer review of this work. Primary Handling Editor: Alireza
Bahadori. A peer review file is available
Reprints and permissions information is available at
http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution-NonCommercial-NoDerivatives 4.0 International License,
which permits any non-commercial use, sharing, distribution and
reproduction in any medium or format, as long as you give appropriate
credit to the original author(s) and the source, provide a link to the Creative
Commons licence, and indicate if you modified the licensed material. You
do not have permission under this licence to share adapted material
derived from this article or parts of it. The images or other third party
material in this article are included in the article’s Creative Commons
licence, unless indicated otherwise in a credit line to thematerial. If material
is not included in thearticle’sCreativeCommons licenceandyour intended
use is not permitted by statutory regulation or exceeds the permitted use,
you will need to obtain permission directly from the copyright holder. To
view a copy of this licence, visit http://creativecommons.org/licenses/by-
nc-nd/4.0/.
© The Author(s) 2024
https://doi.org/10.1038/s43247-024-01895-6 Article
Communications Earth & Environment |           (2024) 5:747 10