Supplementary Information for A long-term survey unveils strong seasonal patterns in the airborne microbiome coupled to general and regional atmospheric circulations Joan Cáliz$ (orcid.org/0000-0001-5714-7431), Xavier Triadó-Margarit$ (orcid.org/0000-0001-5339-3221), Lluís Camarero (orcid.org/0000-0003-4271-8988) and Emilio O. Casamayor (orcid.org/0000-0001-7074-3318) $ equal contribution Emilio O. Casamayor Email:
[email protected] This PDF file includes: Supplementary text Figs. S1 to S8 Table S1 References for SI reference citations
1 www.pnas.org/cgi/doi/10.1073/pnas.1812826115
Supplementary text Bioaerosol collection Atmospheric wet precipitation (both rain and snow) was collected with an automatic wet and dry passive sampler MTX ARS 1010 (MTX, Bologna, Italy) equipped with two 667 cm2 area polyethylene collection buckets and a hygroscopic sensor cell (1). The wet collector remained covered preventing atmospheric inputs until the hygroscopic sensor was activated. Particles deposited in the wet container (i.e. those washed from the atmosphere) were retained onto precombusted (450 °C, 4 hours) Whatman GF/F filters and then dried in a laboratory oven for 4 hours and kept in a dark and dry place (1). For the molecular analysis, the variable V4 and V5 regions of the 16S rRNA gene (~250 nt) were amplified with the primers F515 (5’–GTGCCAGCMGCCGCGGTAA–3’) and R806 (5’– GGACTACHVGGGTWTCTAAT–3’) (2). For the eukaryotic V9 region of the 18S rRNA gene (~150 nt) we used primers 1391F (5’-GTACACACCGCCCGTC-3’) and EukBr (5’-TGATCCTTCTGCAGGTTCACCTAC-3’) (3), following the Earth Microbiome Project (EMP) protocols (www.earthmicrobiome.org). For determination of the chemical composition of the water samples (rain/snow), major cations (Na+, K+, Ca2+, Mg2+ and NH4+) and anions (Cl-, NO3- and SO42-) were analyzed by capillary electrophoresis (Quanta 4000, Waters). Water pH was determined using a low ionic strength electrode (Crison). Acid neutralizing capacity (ANC) was measured by potentiometric Gran titration. Dissolved inorganic (DIC) and organic (DOC) carbon were determined by conversion to CO2 by acidification and catalytic combustion (DOC) and IR spectrometry to measure the CO2 produced, with a Shimadzu TOC-5000 analyzer. Sequence processing Raw rRNA genes sequences were processed using the UPARSE pipeline (4). After merging of read pairs, filtering by read length and with an expected error of 0.25, ~42% of the original reads were retained (i.e., 5,622,407 reads for 16S rRNA gene and 6,417,502 for eukaryotes). The reads were dereplicated and clustered into operational taxonomic units (OTUs) at cut-off 0.03% identity after chimera removal (UCHIME) and excluding the singletons. More than 92% of the globally trimmed, quality filtered sequence pool was mapped back into OTUs. A total of 4,164 prokaryote and 8,248 eukaryote OTUs were obtained and taxonomically assigned with SILVA_119 (5). Chloroplast, mitochondria, Metazoa, Embryophyta (mostly pollen), and unclassified reads were excluded for further analyses. In order to minimize biased effects for differences in sampling effort, the original OTU table was average rarefied (100 random subsamplings, (6)) and set to a depth of 12,500 prokaryotes and 10,000 eukaryotes sequences per sample. Environmental descriptive terms were extracted from closest matches (99% identity) using the SEQenv pipeline (7) for the most abundant OTUs using ENVO terms (8). Sixty-five and fifty-five per cent of the average abundance of prokaryotes and eukaryotes, respectively, were annotated to ENVO terms. It is important here to remark that predicted sources have to be interpreted with caution as annotations rely on homology searches, as well as available source entries, against sequences of from public databases. 2
Statistical Analyses
Factor Analysis (FA) is a dimension reduction technique that substantially reduces the number of original variables, and only a few factors are extracted from the data set. The factors are linear combinations of the original variables, and are computed in order to both be uncorrelated among each other, and be strongly correlated to different groups of input variables. Factors can therefore be interpreted as underlying processes driving the original variables distribution.
3
A
B
Cluster means - Standard
243 backward trajectories
Cluster means - Standard
139 backward trajectories GDAS Meteorological Data
t
-15
o
15
0
NW (7%)
o
s
n
n
NW (7%)
-50 )o n NW fast (4%) -40
at 42.55 N
60
s
-30
0.89 E
NW fast (6%)t -45
at 42.55 N
0.89 E
GDAS Meteorological Data
t
o
sn
t s
n
s
l
n
m l
m
t
n
o
Regional ( 29%)s t
l s
s
t
l
H l
t
s t SE (7%)
s
s l n
H
m o SE (19%) lm m l
n
o
s
m
s
t s ts l t
m
10
t 0
n
Regional (41%)s
m 40
o
-10 t
l
n
45
ts
s
-20 m
Source
Source
t
o
m
o
W (9%)
o N (17%)
W (17%)
N (6%) t
50 -30
t
SW (14%) o 30
l
mm l
n
S (17%)
C
D
Cluster means - Standard
136 backward trajectories
Cluster means - Standard
144 backward trajectories
GDAS Meteorological Data
GDAS Meteorological Data N (12%)t 60
0.89 E
N (7%) t
l
s
m s
-40 NW (20%) t
l
t
-20
t
Source
l
H
ml l Regional (26%) l nmm
W (13%)o
n
o
n
n
s
t
s
t s SE (8%) t
t
-50
l
l
s
m 50
-40
t
l
-30
m
m
-20 l
s
o
NW (15%) m
s
s
t
20
0
50 m
s
s
-60
NW fastm(10%)
t
at 42.55 N
-60
Source
at 42.55 N
0.89 E
NW fastm(7%)
s Regional -10 t (17%) 0 l s
ts ts
W (17%)o 40 m
W fast (10%) o
l
n
n
o
n
o
n
on
n
t
o
SW (18%) o 30 o n
n
30
H
s
o
s
t SW (19%) t s
Fig. S1. Cluster centroids and relative frequency of back-trajectories associated to each cluster during period from 2007 to 2013, analyzed by season: spring (A), summer (B), autumn (C) and winter (D).
4
Factor1
Factor2
Factor3
2.5 Spring Summer 0.0
Autumn Winter
2.5
Fig. S2. Seasonal variation of factors described by the chemical composition of wet
depositions: factor one (NO3-, SO42-, NH4+, DOC and Mg2+), factor two (pH, ANC, DIC, Ca2+ and Mg2+) and factor three (Cl- and Na+).
5
Bray-Curtis distance
0.8
0.6
B 0.8
Bray−Curtis distance
A
0.6
Spring Summer Autumn Winter
0.4 0.4 0.2
Fig. S3. Intra-seasonal community dissimilarity (Bray Curtis index) for bacteria (A) and
eukaryotes (B).
6
A
7
●
500
●
● ●
●
300
●
●
200
4
Spring Summer Autumn Winter
20
3
100
B
5
PD
Shannon
Richness
400
30
6
●
10
2
1200
8
50
900
40
●
600
6
PD
Shannon
Richness
●
4
2
8
●
● ●
600 300
● ●● ● ●● ● ●● ●● ● ●● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ●●●●●●●●● ● ●● ● ● ● ● ●● ● ● ● ● ● ●●● ● ● ● ●
100
● ●
●
●
●
●
● ● ●
● ● ●
●
Shannon (Eukarya)
Richness (Eukarya)
●
● ● ●● ●
●
●
● ●●
200
300
400
Richness (Bacteria)
500
● ● ●
●
4
●
●
●
●
2
●
● ●
●
6
●
● ●
● ●
● ●
●
●
●● ●
●
●
● ● ● ●● ● ● ●
●
●
● ● ●
● ●
● ●
●
5
Shannon (Bacteria)
40
6
7
● ●● ●
30 20 10
●
●
4
●● ●●
●
● ●● ●
● ●● ● ● ●
3
●
●
●
●
●
2
●
●
● ● ● ● ● ●● ● ● ●● ● ● ● ●●● ●● ● ● ●● ●● ● ● ●●●● ● ● ● ● ● ● ●● ●● ● ● ●● ●
50
●
●
● ●
●
900
10
● ●
PD (Eukarya)
1200
Spring Summer Autumn Winter
30 20
300
C
●
● ●
●
● ●
● ● ●
● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ●● ● ● ● ●● ●●● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●●
10
20
●
●
● ● ●
● ● ●
● ●
● ●● ●
● ●
Spring Summer Autumn Winter
●● ●
PD (Bacteria)
30
Fig. S4. Seasonal comparison of alpha diversity indices (richness, Shannon and Faith’s
phylogenetic diversity – PD) for airborne bacteria (A) and eukaryotes (B). Seasonal pairwise comparisons of alpha diversity indices between bacteria and eukaryotes (fitted linear model regression in color lines) (C).
7
Methylobacterium Sphingomonas Massilia
10.00
Sphingomonadales
5.00
Pseudomonas Acidiphilium
Polaromonas Ramlibacter
Acetobacteraceae Mucilaginibacter
Hymenobacter
1.00
Noviherbaspirillum
Rhizobiales
0.10
Mean relative abundance where present (%)
0.05
High Rank Taxonomic Group Alphaproteobacteria Betaproteobacteria
Gammaproteobacteria Bacteroidetes
10.00
50
100 Agaricomycetes Tremellales
Chytridiomycetes
Chytridiomycetes
5.00
Pleosporales
Tremellomycetes Pezizomycotina Ciliophora;Colpodea Fungi Agaricomycotina Cercozoa;Thecofilosea Trebouxiophyceae Capnodiales Taphrina Polyporales Cercozoa Fungi;Dikarya Pucciniomycotina
1.00
Helotiales
Chaetothyriales Lecanorales
Dothideales Agaricales
0.10 0.05
0.01
High Rank Taxonomic Group Archaeplastida; Chlorophyta Opisthokonta; Fungi SAR; Alveolata SAR; Rhizaria
50
Number of samples
100
Fig. S5. Mean relative abundances (where present) and occupancy relationships for
bacterial and eukaryal taxa (upper and lower panel, respectively).
8
B
Acidobacterales
Alteromonadales Bacillales Bdellovibrionales Burkholderiales Caulobacterales Corynebacterales Cytophagales
Opistokonta; Fungi Chytridiomycota Ascomycota
A
Spizellomycetales Insertae Sedis - other1 Insertae Sedis - other2 Dothideales Pleosporales Dothideomycetes-other Chaetothyriales Lecanorales Agaricales
Deinococcales
Boletales
Enterobacteriales
Spring
40
Relative abundance (%)
30
20
10
Xanthomonadales
Summer
Autumn
Tetramitia Colpodea
20
Sphingomonadales
Gracilipodida Trebouxiophyceae
15
Sphingobacteriales
Basidiomycota - other
10
Rhizobiales Rhodospirillales
Tremellales Ustilaginales
5
Oceanospirillales
Pucciniales
0
Frankiales Myxococcales
Amebozoa Chlorophtyta Heterolobosea Alveolata Basidiomycota
Agaricomycetes - other
Relative abundance (%)
Winter
Fig. S6. (A) Bacterial orders and (B) Eukaryal taxa (when possible, at the order level) with significantly different relative abundances among seasons. Minor groups, i.e., relative abundance < 1%, are not shown.
9
Percentage identity of OTU sequences to nt database
100
95
90
85
Bacteria Eukarya
80 25
50
75
Occurrence (Number of samples)
100
125
Fig. S7. OTU sequence resemblance to databases, based on its percent identity values, and
occupancy relationships. Data from both bacterial and eukaryal OTUs are shown.
10
A
Aquatic
Freshwater
Cropland
Urban
Terrestrial
Frequency
600 400 200 Season Spring
Forest
Marine
Mediterranean sea
Desert
Mangrove
Frequency
400 200
1500 1000 Frequency
Autumn Winter
600
B
Summer
500
Aquatic
Freshwater
Forest
Tropical moist broadleaf forest
Marine Season Spring Summer Autumn Winter
Fig. S8. Seasonal distribution of depositions according to the predicted source of biomes
related to airborne bacteria (A) and eukaryotes (B). Biomes are ordered from left to right according to seasonal prevalence. Note that because the skewed distribution, the y-axis is squared.
11
Number of days with Saharan dust intrusion per month
20
15
Spring Summer
10
Autumn Winter
5
0
Fig. S9. Comparison of the occurrence of Saharan dust intrusions per month among
seasons. Data retrieved from www.calima.ws
12
Table S1. Correlation coefficients of the three factors extracted by Factor Analysis
Factor1 NO3-
0.924
2-
0.772
+
0.687 0.606 0.600
SO4
NH4 DOC Mg2+ pH ANC DIC Ca2+ ClNa+
Factor2
Factor3
0.604 0.831 0.802 0.770 0.733 0.925 0.851
13
References for SI reference citations 1. 2. 3. 4. 5. 6. 7. 8.
Hervas A, Camarero L, Reche I, & Casamayor EO (2009) Viability and potential for immigration of airborne bacteria from Africa that reach high mountain lakes in Europe. Environ Microbiol 11(6):1612-1623. Caporaso JG, et al. (2011) Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc Natl Acad Sci U S A 108 Suppl 1:45164522. Stoeck T, et al. (2010) Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water. Mol Ecol 19(s1):21-31. Edgar RC (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Methods 10(10):996-998. Quast C, et al. (2013) The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41(Database issue):D590-596. Caliz J, et al. (2015) Influence of edaphic, climatic, and agronomic factors on the composition and abundance of nitrifying microorganisms in the rhizosphere of commercial olive crops. PLoS One 10(5):e0125787. Sinclair L, et al. (2016) Seqenv: linking sequences to environments through text mining. PeerJ 4:e2690. Buttigieg PL, Morrison N, Smith B, Mungall CJ, & Lewis SE (2013) The environment ontology: contextualising biological and biomedical entities. J Biomed Semant 4(1):1-9.
14
