Content uploaded by Carsten Ambelas Skjøth
Author content
All content in this area was uploaded by Carsten Ambelas Skjøth
Content may be subject to copyright.
ORIGINAL PAPER
High bacterial 16S rRNA gene diversity
above the atmospheric boundary layer
Ulla Li Zweifel •A
˚ke Hagstro
¨m•Karin Holmfeldt •
Runar Thyrhaug •Camilla Geels •Lise Marie Frohn •
Carsten A. Skjøth •Ulrich Gosewinkel Karlson
Received: 16 August 2011 / Accepted: 25 January 2012 / Published online: 1 March 2012
ÓSpringer Science+Business Media B.V. 2012
Abstract The atmosphere is host to an omnipresent
bacterial community that may influence fundamental
atmospheric processes such as cloud formation and
precipitation onset. Knowledge of this bacterial com-
munity is scarce, particularly in air masses relevant to
cloud formation. Using a light aircraft, we sampled
above the atmospheric boundary layer—that is, at
heights at which cloud condensation occurs—over
coastal areas of Sweden and Denmark in summer
2009. Enumeration indicated total bacterial numbers
of 4 910
1
to 1.8 910
3
m
-3
air and colony-forming
units of 0–6 bacteria m
-3
air. 16S rRNA gene libraries
constructed from samples collected above the Baltic
Sea coast revealed a highly diverse bacterial commu-
nity dominated by species belonging to the genera
Sphingomonas and Pseudomonas. Bacterial species
known to carry ice-nucleating proteins were found in
several samples. Modeled back trajectories suggested
the potential sources of the sampled bacteria to be
diverse geographic regions, including both marine and
terrestrial environments in the northern hemisphere.
Several samples contained 16S rRNA genes from
plant chloroplasts, confirming a terrestrial contribu-
tion to these samples. Interestingly, the airborne
bacterial community displayed an apparent seasonal
succession that we tentatively ascribe to in situ
succession in the atmosphere.
Keywords Biological aerosols Airborne
microorganisms Trajectories Cloud formation
Ice nucleation
1 Introduction
Microorganisms are part of the airborne suspended
particles and may affect fundamental atmospheric
processes such as cloud formation and ice crystal
Runar Thyrhaug: opus posthum.
Electronic supplementary material The online version of
this article (doi:10.1007/s10453-012-9250-6) contains
supplementary material, which is available to authorized users.
U. L. Zweifel (&)
Department of Cell and Molecular Biology, University
of Gothenburg, Box 462, 405 30 Go
¨teborg, Sweden
e-mail: ullali.zweifel@gu.se
A
˚. Hagstro
¨mK. Holmfeldt
School of Natural Sciences, Linnaeus University, 391 82
Kalmar, Sweden
Present Address:
K. Holmfeldt
Ecology and Evolutionary Biology Department,
University of Arizona, Tucson, AZ 85721, USA
R. Thyrhaug
Department of Biology, University of Bergen, PO Box
7803, 5020 Bergen, Norway
C. Geels L. M. Frohn C. A. Skjøth U. G. Karlson
Department of Environmental Science, Aarhus
University, PO Box 358, 4000 Roskilde, Denmark
123
Aerobiologia (2012) 28:481–498
DOI 10.1007/s10453-012-9250-6
growth (Christner et al. 2008a;Mo
¨hler et al. 2007; Sun
and Ariya 2006). In terms of number, microorganisms
constitute a minor fraction of the total aerosols,
typically \1 parts per thousand (Martin et al. 1997;
Zhou et al. 2002). However, due to their physical
properties, biological particles may affect the meteo-
rological processes much more than their numbers
would imply. For example, due to their relatively large
size, bacteria and other microorganisms are expected
to contribute to the formation of precipitation by
acting as condensation nuclei (Pruppacher and Klett
1997; Franc and DeMott 1998). However, it has been
proposed that the suitability of bacteria for acting as
cloud condensation nuclei is species specific, as their
wettability depends on the chemical composition,
structure, and hydrophilicity of the outer surface of the
bacterial cell wall (reviewed by Sun and Ariya 2006)
Microorganisms can also act as nuclei in the transfor-
mation of water vapor into small ice crystals, which
grow by attracting additional water vapor and by
scavenging super-cooled water droplets when they fall
through the cloud (Mo
¨hler et al. 2007). Furthermore,
some bacteria have ice nucleation proteins (INPs) on
the cell surface that can raise the temperature for the
transition from water vapor to ice up to -2°C, thus
contributing to the formation of ice crystals (Kawa-
hara 2002). Bacterial species known to carry genes
encoding for INP belong to, for example, Pseudomo-
nas sp., Xanthomonas sp., and Erwinia sp. (Kobash-
igawa et al. 2005; Muryoi et al. 2003). Bacteria and
their role in ice nucleation have recently been
emphasized by findings that up to 85% of ice
nucleation activity (INA) in precipitation (snow)
samples is lost when the samples are treated with
lysozyme, that is, upon degradation of bacterial cell
walls (Christner et al. 2008b). This underscores that
species-specific properties, and the characteristics of
the airborne bacterial community are highly relevant
to atmospheric processes.
Knowledge of bacteria in the atmosphere has
traditionally been based on cultivation-dependent
techniques (Di Giorgio et al. 1996; Fang et al. 2007;
Kaarakainen et al. 2008; Kellogg and Griffin 2006;
Shaffer and Lighthart 1996), leaving uncertainty about
the representativeness of these species in the natural
environment, since many bacterial strains cannot grow
on nutrient-rich agar plates (Staley and Konopka
1985; Simu et al. 2005). This uncertainty has recently
been reduced by using cultivation-independent
sampling techniques and identification of the 16S
rRNA gene. These studies have revealed that the
airborne community is highly diverse, including
representatives of well-known bacterial genera, the
density of which in the atmosphere varies with time
and location (Bowers et al. 2009; Brodie et al. 2007;
Despres et al. 2007; Fahlgren et al. 2010; Maron et al.
2005; Cho and Hwang 2011). It should be noted that
samples of airborne bacteria have mainly been
collected relatively near the ground, and occasionally
at high-altitude mountain sites (Amato et al. 2007;
Bowers et al. 2009; Despres et al. 2007; Herva
`s et al.
2009; Herva
`s and Casamayor 2009; Reche et al.
2009). However, for bacteria to act in processes such
as cloud condensation and ice nucleation, they have to
prevail and persist at altitudes at which these processes
occur.
Although clouds can form through several pro-
cesses and at various altitudes, a crucial altitude is
above the atmospheric boundary layer (ABL). The
ABL is the lowest part of the troposphere and is
directly influenced by the structure of the earth’s
surface and rising thermals of air, which originate at
the surface. The height of the ABL varies with time of
day, season, and geographical location, typically from
a few hundred meters at night to between 1,000 and
2,000 m in the day (Smith et al. 2008). Globally,
condensation succeeded by precipitation occurs
mostly in convective clouds formed at the top of the
ABL. In temperate regions, however, convective
clouds usually remain too small (e.g., cumulus clouds)
to produce precipitation and only occasionally reach
rain shower size (e.g., cumulonimbus). Instead, the
annual precipitation in temperate regions stems from
clouds initiated above the ABL through production of
stratiform clouds.
In this study, we have sampled the airborne
bacterial community using light aircraft in two
independent sampling campaigns: one conducted
above the ABL over the Baltic Sea coast, Sweden,
and one conducted above and within the ABL over
land and coastal waters of Denmark. In the Danish
campaign, samples were analyzed for the number of
colony-forming units (CFU) of bacteria, while in the
Swedish campaign, samples were analyzed for total
number of bacteria as counted by microscopy and
molecular analysis of bacterial 16S rRNA gene
sequences. Thus, this study provides the number of
bacteria based on both cultivation-dependent and
482 Aerobiologia (2012) 28:481–498
123
-independent counting, and a culture-independent
analysis of the composition of the airborne community
in air masses relevant to cloud formation in a
temperate region, where the history of the air masses
was inferred using modeled back trajectories.
2 Materials and methods
2.1 Air sampling
Sampling of airborne bacteria was carried out using a
light aircraft. A Diamond Katana DA-20 with a built-
in NACA duct air inlet was used as the sampling
vessel. Before each flight, the altitude of the ABL was
predicted using the operational THOR weather and air
pollution forecast system (Brandt et al. 2001a,b). In
addition, during the flight, the ABL was visually
inspected by observing layers of haze. After takeoff,
sampling was initiated and terminated 200 m above
ground for sampling within the ABL, and at the
appropriate above ABL altitude for sampling above
the ABL.
Swedish samples: Sampling was conducted six
times over the July 14–August 9, 2009, period near the
Island of O
¨land, Baltic Sea, Sweden, at approximately
500 m above sea level (a.s.l). The samples were
collected between 08:00 and 10:00 h while the
airplane was flying a transect from 20 km to the south
and 20 km to the north of 56.95 N 016.93 E. All
samples were collected above the ABL and further
processed by means of molecular analysis and bacte-
rial enumeration.
Danish samples: A parallel sampling campaign
using an identical sampling system was conducted
over land and sea areas of Denmark. Ten samples were
collected over the July 15–August 25, 2009, period in
the area between the southern tip of Lolland (54.57 N
011.84 E) and the northern tip of Læsø (57.35 N
010.90 E), approximately 670 km west of the Swedish
transect. Individual flights covered distances of up to
300 km in the north–south and 50 km in east–west
direction. For comparison of the bacterial numbers,
and as a quality control of the sampling system, five
samples were taken above the ABL at altitudes
between 500 and 2,000 m a.s.l., and five within the
ABL. During one flight, using two independent
sampling systems, sampling was done both above
and below the appropriate altitudes, which resulted in
two separate samples, one from above and one from
within the ABL, both taken on the same day and at the
same location. The duration of individual samplings
was between 1.0 and 3.5 h per sample. Samples were
processed for the enumeration of bacterial CFUs.
Sample collection: All air samples were collected
using the inner water-holding vortex chamber of the
Ka
¨rcher DS 5600 commercial vacuum cleaner (Alfred
Ka
¨rcher GmbH & Co. KG, Germany). The vortex
chamber is designed to form a water vortex causing the
particles carried by the air stream to be captured
(scrubbed) in the liquid contained in the vortex
chamber. The vortex chamber was connected to the
NACA duct inlet, and a water vortex was created by
ram air, that is, the air pressure created by aircraft
motion. The NACA duct air inlet and fittings were
made of stainless steel, and the tubes connecting the
NACA duct inlet to the vortex chamber were made of
aluminum-coated cardboard to prevent static interac-
tions (as would be caused, for example, by plastic
material). In this mode, the airflow through the vortex
chamber was kept at 1.8 m
3
min
-1
, measured during
operation by fitting a wind speed meter with a rotating
vane sensor (Kimo LVB) to the modified outlet pipe of
the vortex chamber. The HEPA filter from the original
vacuum cleaner was removed permanently and
replaced with a stainless steel plate, holding a curved
stainless steel pipe serving as the outlet. Prior to
sampling, the vortex chamber and all material in
contact with the sample were washed in 1 M HCl,
rinsed in filtered (0.2-lm pore size, Supor filter, Pall
Corporation, East Hills, New York) ultrapure water
(Milli-Q plus 185, Millipore, Billerica, MA), and then
sealed and stored in sterile polycarbonate bags.
Immediately before sampling, 500 mL of freshly
filtered (as above) 19PBS was used to rinse the
vortex chamber. The vortex chamber was immediately
refilled with 2 L of freshly filtered 19PBS that served
as collection buffer, and the whole collection system,
with inlet and outlet pipes sealed, was installed in the
aircraft. Sampling was initiated in flight by opening
inlet and outlet without exposing them to cockpit air.
The inside of the connector and the flexible tube was
sprayed with ethanol (70%) when fixed to the NACA
air inlet. Prior to sampling (i.e., while climbing to
altitude) air was flowing free through the pipe into the
cockpit thus removing all traces of the ethanol. A
plastic bag loosely fitted over the end of the tube was
removed before connecting the flexible pipe to the
Aerobiologia (2012) 28:481–498 483
123
sampler, and vinyl gloves were used during handling
to avoid contamination. Sampling was terminated
during flight by disconnecting the system and closing
its openings in reverse order. Over 1–4 h of sampling,
up to 300 mL of buffer was lost due to evaporation,
which was within the design limits of the vortex
chamber. After sampling, the remaining buffer was
further processed as described below.
2.2 Bacterial enumeration
Swedish samples: The total number of bacteria was
estimated from a 50-mL subsample of the collected
buffer by fixing with 4% (v/v) formalin, followed by
staining with DAPI (4,6-diamidino-2-phenylindole,
Sigma-Aldrich, St. Louis, MO) (Porter and Feig 1980)
at a final concentration of 0.5 lg DAPI mL
-1
sample
for 10 min. Stained samples were filtered onto black
0.2-lm pore size polycarbonate filters, and bacterial
cells were counted using epifluorescence microscopy
(EFM). The total number of cells in 50 counting fields
at 5009magnification was used to calculate bacterial
numbers. The theoretical detection limit was calcu-
lated to be 2 cells m
-3
of air. As a control, the liquid
that was used to rinse the vortex chamber before
sampling was also processed for bacterial enumeration
by microscopy and growth on ZoBell agar plates
(ZoBell 1946) to check for possible contamination of
the sampling system.
Danish samples: The complete sample was filtered
onto a 0.2-lm pore size, 47-mm-diameter Supor filter
(Pall Corporation). The filter was then immediately
vortexed in 2 mL of freshly filtered PBS. The number
of CFUs was estimated by plating 100, 10, and 1 lLof
filter extract onto ZoBell agar plates. Colonies were
counted after 10 days’ incubation at 30°C. The
theoretical detection limit was calculated to be
1 CFU 2–20 m
-3
of air, depending on sampling time.
2.3 DNA extraction, 16S rRNA gene
amplification, and cloning
Molecular analysis was conducted on Swedish sam-
ples only. After sampling for bacterial enumeration,
the residual collection buffer was filtered onto a 0.2-
lm pore size, 47-mm-diameter Supor filter (Pall
Corporation) and kept at -80°C until further analysis.
On one occasion, the PBS used to rinse the vortex
chamber was also filtered onto a Supor filter, as
described for the sample, and subjected to molecular
analysis. DNA extraction was performed using a
CTAB-based protocol as described by Fahlgren et al.
(2010), except that the extracted filter was removed
from the procedure before adding chloroform/isoamyl
alcohol. The 16S rRNA gene was amplified using the
universal bacterial primers 27F (50-AGAGTTTGAT
CMTGGCTCAG-30) and 1492R (50-TACGGYTACC
TTGTTACGACTT-30) (Invitrogen, Carlsbad, CA).
The PCR was performed using a final concentration of
0.2 lM of each primer and the illustra PuReTaq
Ready-To-Go PCR Beads (GE Healthcare/Life Sci-
ences, Uppsala, Sweden) according to the manufac-
turer’s instructions. PCR cycling conditions were as
follows: initial denaturation at 95°C for 2 min fol-
lowed by 30 cycles consisting of 95°C for 30 s, 50°C
for 30 s, 72°C for 45 s, and a final elongation at 72°C
for 7 min. The PCR products were purified using the
E.Z.N.A. Cycle-Pure Kit (Omega Bio-Tek, Norcross,
GA) and cloned using the TOPO TA Cloning Kit
(Invitrogen). All sequencing work was done by
Swegene, Lund, Sweden, using the chain-termination
method and the universal bacterial primer 27F (Invit-
rogen). A blank filter was always run parallel to the
samples throughout the extraction process and PCR
amplification.
2.4 Sequence analysis
The 16S rRNA gene sequences were analyzed and
edited using the program 4Peaks 1.7.2 (developed by
A. Griekspoor and T. Groothuis and available at http://
www.mekentosj.com). The resulting sequences were
250–660 bp in length, most being approximately
600 bp in length. The nucleotide sequences were
deposited in GenBank with the accession numbers
JF268792–JF269170 and JF278091–JF278167. Pos-
sible chimera formation was checked using the pro-
gram Bellerophon (Huber et al. 2004). Sequences
indicated as possible chimeras were further analyzed
against the GenBank sequence with the highest simi-
larity using the program Pintail (Ashelford et al. 2005);
available at http://www.bioinformatics-toolkit.org/
Pintail/). Independent of the results from Bellero-
phon, all sequences with an identity of \95% to the
highest BLAST hit (Altschul et al. 1990) were also
checked using Pintail. A total of 11 sequences were
removed through this process. The remaining sequen-
ces were assembled into consensus sequences based on
484 Aerobiologia (2012) 28:481–498
123
[97% sequence identity using SeqMan II (Lasergene
v. 7, DNASTAR, Madison, WI). The consensus
sequence was considered to represent one operational
taxonomic unit (OTU). The consensus sequences were
analyzed using BLAST and labeled according to a
comprehensive overview of the sequences with highest
similarity in GenBank. In order to avoid mislabeling of
the OTUs, we used the following stringent criteria:
\90% identity was labeled as unknown bacteria,
91–94% identity according to phylum, 95–99% iden-
tity according to genus, and 100% identity to the spe-
cies level. From the initial consensus sequence
assembly and BLAST search, it became apparent that
several of 16S rRNA gene sequences belonged to
plants. Since 16S rRNA genes from chloroplasts and
mitochondria are highly conserved, these consensus
sequences were dissolved and reassembled at 99%
identity (Soltis et al. 1998). These sequences could still
originate from several species, so the consensus
sequences were labeled according to a set of stringent
rules: (1) all top-scoring BLAST hits of identical
sequence identity were considered; (2) taxonomic
affiliation was based on the lowest level of shared
taxonomy among the top-scoring hits of identical
sequence identity, unless (3) the top-scoring BLAST
hits were \97% identical to known plant-related 16S
rRNA genes, in which case the taxonomic affiliation
was not assigned further than the level of order; and (4)
subject to the rule that taxonomic affiliation was never
assigned further than the level of family.
Sequences were imported into an ARB database of
ca 52,000 reference sequences and aligned automat-
ically toward their closest relatives using the inte-
grated alignment module within the ARB package
(Ludwig et al. 2004). Backbone trees based on
sequences of C1,400 nucleotides were calculated
using the algorithms neighbor joining, maximum
parsimony, and maximum likelihood to verify the
validity of branching patterns. Sequences \1,400
nucleotides were added to the trees afterward accord-
ing to maximum parsimony criteria, which does not
allow changes in the overall topology. A maximum
likelihood tree (1,000 bootstraps) is presented in this
paper.
2.5 Back trajectories
Back trajectories were calculated for the Swedish
samples. The trajectory calculations were based on
meteorological data from the operational THOR
weather and air pollution forecast system (Brandt
et al. 2001a,b) using the Eta weather forecast model
(Janjic 1990,1994) in combination with a flexible
trajectory model (Skjøth et al. 2002). This trajectory
model has been applied in a number of aerobiological
studies for the analysis of historical data (Kasprzyk
et al. 2011; Sikoparija et al. 2009; Skjøth et al. 2008;
Stach et al. 2007) and for the operational forecasting of
back trajectories 3 days in advance (Hertel et al.
2003). For this work, back trajectories were calculated
by following the matrix-style methodology described
by Skjøth et al. (2007). The back trajectories were run
at hourly intervals throughout the campaign periods.
Each back trajectory was run for 4 days with a 1-h
step.
3 Results
3.1 Bacterial enumerations
During the Swedish sampling campaign, samples for
bacterial enumeration were collected on six occasions
concomitant with samples for molecular analysis. The
total number of particles in each of the six samples was
low, only 0–5 tentative bacterial cells being detected
in each counting grid. The control, that is, a sample of
the buffer used to rinse the vortex chamber before
sampling, contained bacteria-sized fluorescent parti-
cles constituting up to 10% of the total cell count.
Since agar plating yielded no CFUs, these counts were
subtracted from the total number of cells estimated in
the samples. The resulting bacterial numbers varied
between 4 910
1
and 1.8 910
3
bacteria m
-3
of
sampled air (Table 1), but due to the low number of
cells, the standard deviation was up to 250% of the
mean. Cyanobacteria as well as autofluorescent cells
in sizes up to 50 lm were seen occasionally in all
samples. The buffer used to rinse the chamber did not
yield any CFUs.
During the Danish sampling campaign, the num-
bers of CFUs of bacteria were estimated in 10 samples
collected between July 15 and August 9, 2009. These
samples yielded from 0 (1 sample) to 6 CFU m
-3
of
sampled air (Table 2). Within 5–19 days incubation,
between 50 and 97% (average 83.9 ±14.4%) of all
colonies developed a dark yellow, orange or red color,
indicating strong formation of pigments. Also in this
Aerobiologia (2012) 28:481–498 485
123
campaign, the buffer used to rinse the vortex chamber
before sampling was devoid of CFUs.
3.2 Molecular analyses
The samples collected in the air mass overlying the
Baltic Sea were processed using DNA extraction,
followed by 16S rRNA gene amplification, cloning,
and sequencing of the 16S rRNA gene. From each of
the samples, 96 clones were selected and sequenced.
Between 67 and 88 sequences per clone library were
deemed of appropriate quality for further analysis
(Table 1). For each sampling date, the bacterial
sequences were assembled into consensus sequences
based on [97% sequence identity resulting in 17–35
bacterial operational taxonomic units (OTUs) depend-
ing on sampling date. In all samples, \5 OTUs made
up at least 50% of the total sequences, that is, a limited
number of OTUs dominated the clone libraries. Of the
remaining OTUs, most were singletons, that is,
containing only one sequence (Table 1, Supporting
information TS1). The OTUs were distributed over the
Alpha-, Beta-, and Gammaproteobacteria (Fig. 1a),
Firimicutes, Actinobacteria, and Cyanobacteria
(Fig. 1b), and Bacteroidetes and Acidobacteria
(Fig. 1c).
Seen over the sampling period, bacteria from the
Alphaproteobacteria dominated the community, fol-
lowed by Gammaproteobacteria and Bacteroidetes
(Fig. 2). The most frequently occurring sequences
belonged to the bacterial genus Sphingomonas, occur-
ring in all six samples and being among the dominant
bacteria in four of them (Table 3). When comparing
the Sphingomonas-like sequences individually, it was
clear that identical 16S rRNA gene sequences were
found on several different sampling dates, for
Table 1 Characterization of air samples collected above the atmospheric boundary layer in the air mass overlying the Baltic Sea
coast, near O
¨land, Sweden, summer 2009
Jul 14 Jul 16 Jul 22 Aug 2 Aug 8 Sept 9
Bacterial density (EFM counts m
-3
) 288 43 239 1873 89 87
16S rRNA clone library information
Total number of sequences 68 86 74 67 77 85
Number of bacterial OTUs ([97% sequence identity) 17 35 35 28 31 26
Number of prokaryotic singletons 13 15 27 16 23 20
Prokaryote sequences (%) 47 100 99 93 91 69
Plant-related sequences (%) 53 0 1 7 9 31
Potential INA bacteria (% of bacterial sequences) 41 0 0 9.5 1.5 8.5
Table 2 Bacterial density in
air samples collected above and
within the atmospheric
boundary layer over Denmark,
summer 2009
Sampling
date
Air volume
sampled
(m
3
)
Above/within
ABL
Remarks Bacterial density
(CFU m
-3
)±SD
Jul 15 137 Within Over land 1.97 ±0.07
Jul 18 198 85% above Over land 0.03 ±0.03
Aug 3 108 Within Over land, after rain 2.99 ±0.11
Aug 11 212 Above Over land, after rain 1.64 ±0.15
Aug 17 378 Within Mostly over land 3.12 ±0.46
Aug 18 268 Within Mostly over land 1.31 ±0.57
Aug 18 144 Above Over sea 0.00
Aug 20 304 Within 50% Over coastal
waters, 50%
over land
0.37 ±0.02
Aug 24 184 90% Above Over sea 0.75 ±0.05
Aug 25 236 Above Over sea 5.54 ±0.47
486 Aerobiologia (2012) 28:481–498
123
Fig. 1 Phylogenetic tree
(maximum likelihood) of
aAlpha-, Beta-,and
Gammaproteobacteria,
bFirimicutes,
Actinobacteria, and
Cyanobacteria, and
cBacteroidetes,
Acidobacteria, and
Planctomycete. Reference
sequences ([1,400 bp) are
shown in regular font
followed by GenBank
accession numbers. The
OTUs identified in this study
are marked in bold and
followed by sampling
month–date and the
designated number of each
consensus sequence.
Sequence assembly was
based on[97% sequence
identity. The number of
sequences in each consensus
assembly and the accession
number of individual
sequences are provided in
the Supporting information
to this article. The
taxonomic affiliation of the
OTUs was deduced from a
BLAST search and a set of
rules as outlined in
Experimental procedures.
The tree was rooted with 3
Archaea sequences (not
shown). Phylogenetic
relationships were
bootstrapped 1,000 times
and values[50% are shown
Aerobiologia (2012) 28:481–498 487
123
Fig. 1 continued
488 Aerobiologia (2012) 28:481–498
123
Fig. 1 continued
Aerobiologia (2012) 28:481–498 489
123
Table 3 Dominant bacterial operational taxonomic units (OTUs) and rare bacterial OTUs of special interest in samples collected
above the atmospheric boundary layer in the air mass overlying the Swedish Baltic Sea coast, summer 2009
Sampling
date
Dominant bacterial OTUs
([5% of bacterial sequences)
Rare bacterial OTUs of special interest
(\5% of bacterial sequences)
Indicated taxonomy of plant-
related
consensus sequences
(nsequences)
Jul 14 Pseudomonas graminis (41%) Sphingomonas sp. (ubiquitous) Magnoliophyta (25)
Pinaceae (6)
Poaceae (4)
Bryophyta (3)
Oomycetes (2)
Jul 16 Unknown Acidobacterium (14%)
Unknown Cyanobacterium (12%)
Propionibacterium acnes (6%)
Unknown Bacteroidetes bacterium
(6%)
Sphingomonas sp. (ubiquitous)
Pseudomonas sp. (ubiquitous)
Erwinia sp. (plant pathogens)
–
Jul 22 Rahnella sp. (21%)
Sphingomonas sp. (14%)
Unknown Cyanobacterium (10%)
Aurantimonas sp. (previously isolated
from clouds)
Clostridium sp. (obligate anaerobes)
Propionibacterium acnes (common)
Bryophyta (1)
Aug 2 Sphingomonas sp. (24%)
Pedobacter sp. (11%)
Pantoea agglomerans (5%)
Agrobacterium tumefaciens (5%)
Aurantimonas sp. (previously isolated
from clouds)
Clostridium sp. (obligate anaerobes)
Pseudomonas syringae (INA bacterium)
Magnoliophyta (4)
Pinaceae (1)
Aug 8 Sphingomonas sp. (46%)
Pseudomonas sp. (6%)
Aurantimonas sp. (previously isolated
from clouds)
Clostridium sp. (obligate anaerobes)
Erwinia persicina (plant pathogen)
Magnoliophyta (5)
Bryophyta (1)
Oomycetes (1)
Aug 9 Sphingomonas sp. (41%)
Pseudomonas sp. (12%)
Pantoea agglomerans (7%)
Clostridium sp. (obligate anaerobes)
Propionibacterium acnes (common)
Magnoliophyta (22)
Bryophyta (3)
Marchantiophyta (1)
0
20
40
60
80
100
% of sequences
Sampling date
Alphaproteobacteria
Bacteroidetes
Betaproteobacteria
Cyanobacteria
Actinobacteria
Firmicutes
Acidobacteria
Gammaproteobacteria
Plant related sequences
Spinghomonas sp.
14 Jul 16 Jul 22 Jul 2 Aug 8 Aug 9 Aug
Fig. 2 Frequency of
bacterial phyla and plant-
related sequences versus
time in samples collected
above the atmospheric
boundary layer over the
Baltic Sea coast, Sweden,
summer 2009
490 Aerobiologia (2012) 28:481–498
123
example, on the July 22, August 8, and August 9 (data
not shown). Pseudomonas spp. was also frequently
occurring and found to be dominant in three of the
samples. Pantoea agglomerans and Propionibacte-
rium acnes were two species that appeared in
relatively low abundance (few sequences) but on
several sampling occasions. On July 16, the dominant
OTUs displayed low resemblance to existing
sequences in GenBank. For example, the abundant
cyanobacterial sequences retrieved on this date were
most similar to the genera Calothrix and Rivularia,
although they were only 92% similar to previously
reported sequences. Bacteria with high 16S rRNA
similarity (C99%) to species known to possess INA
were found on four sampling dates, making up 1–40%
of the bacterial 16S rRNA sequences in these samples
(Tables 1,2).
Plant-related 16S rRNA gene sequences were
included in the clone libraries on five occasions. Most
of these (78%) were closest related to 16S rRNA genes
from chloroplasts, and the remaining were most
similar to mitochondria. On July 14, when the sample
was dominated by plant-related sequences, the con-
sensus sequences indicated the presence of chloro-
plasts and mitochondria from disparate groups such as
Magnoliophyta (flowering plants), Pinaceae (pine
family), Poaceae (grass family), and Bryophyta
(ferns). Single sequences related to Marchantiophyta
(liverworts) and Oomycetes (water moulds) were
found occasionally (Table 3, Supporting information
Fig. S1, Supporting information TS2).
When viewing the occurrence of major bacterial
phyla and plant-related sequences over time (Fig. 2),
the sample retrieved on July 14 was clearly dominated
by sequences from chloroplasts and mitochondria and
the Gammaproteobacteria. The sample collected on
July 16 displayed a different profile, being dominated
by Cyanobacteria,Acidobacteria, and the gram-
positive phyla Actinobacteria and Firmicutes. There-
after, followed three sampling dates when the Alpha-
proteobacteria and Bacteroidetes were included
among the dominant OTUs; finally, on August 9,
16S rRNA sequences of chloroplast origin were once
again predominant (Fig. 2).
A critical aspect of air sampling is the low
abundance of cells—hence, the importance of elimi-
nating and controlling the risk of contamination. To
check for DNA contamination, we analyzed the buffer
that was used to rinse the vortex chamber immediately
before each sampling, finding no detectable PCR
product. In addition, blank filters and reagents that
were analyzed in parallel to samples never yielded any
PCR product.
3.3 Trajectory analyses
In the ABL, physical factors generate turbulence that
results in local mixing of the air mass, which in turn
means that the air mass in the ABL carries particles
originating from local sources. Roughly, the ABL is
about 1,000 m (Seinfeld and Pandis 1998), with large
variations during the day, for example, from a few
hundred meters at night to between 1,000 and 2,000 m
in the day (Smith et al. 2008). In contrast, the air above
the ABL is not directly affected by local sources, and
most particles found therein have been carried by
long-distance transport. Vertical mixing in the ABL is
also affected by atmospheric stability. Under stable
conditions, atmospheric physics suppress or may even
prevent vertical mixing of air, thus restricting the
dispersion of air pollution (Hertel et al. 2009). The
Swedish samples were collected above the ABL and
during morning hours (08:00–10:00 h local time)
under stable weather conditions. Both factors can on
their own suppress or totally prevent dispersion of
local pollutions to the measuring height. It is therefore
safe to discount contribution from local sources, when
evaluating the source region.
The backward trajectories revealed six distinct
source regions (Table 4, Fig. 3). On the first three
sampling dates, the sampled air mass had passed over
forested areas in the previous 12 h before sampling,
while on the three last sampling dates the air mass had
passed over sea areas the previous 12 h. These last
12 h are mainly nighttime values (20:00 h–08:00 h)
reflecting limited convection due to lack of sunshine.
The 12–36-h period includes nighttime and daytime
values. On the first four sampling dates, the air masses
had passed over sea and over large agricultural areas,
devoted to grass and cereal production, in Denmark,
Germany, and Poland. On the last two sampling dates,
the air masses had passed over land areas in the Baltic
region that contain large forested areas. When con-
sidering the source areas 36–96 h back in time, the
trajectories passed mainly over sea on the first, third,
and fourth sampling occasions, while on the second,
fifth, and sixth they passed mainly over land areas.
Considered over 4 days, the sampled air mass had
Aerobiologia (2012) 28:481–498 491
123
passed over both land and sea areas, and the trajectory
matrix indicated limited spread among the trajectories
on all sampling occasions.
4 Discussion
This study presents the number and composition of the
airborne bacterial community above the ABL in
samples collected along the Swedish Baltic Sea coast
and above Denmark. At this altitude, clouds are
formed from the moisture contained in the air mass
overlying the ABL. Thus, the bacteria in this air mass
may contribute to condensation (cloud formation) and
ice crystal formation (precipitation).
The number of bacteria as estimated by CFUs
during the Danish sampling campaign was only
0–6 CFU m
-3
of sampled air. The number of bacterial
CFUs in air samples collected near the ground is
typically higher, ranging from a few CFU up to
1910
4
m
-3
(Cho and Hwang 2011; Lighthart 1997;
Mancinelli and Shulls 1978). Likewise, the number of
bacteria estimated using microscopy during the Swed-
ish sampling campaign was 4 910
1
to 1.8 910
3
bac-
teria m
-3
, while the total number of airborne bacteria
near the ground or water sources is typically in the
range of 5 910
3
to 1 910
5
bacteria m
-3
(Bauer et al.
2002; Cho and Hwang 2011). The lower number of
airborne bacteria at higher altitudes can be explained
by the dilution effect of distance from potential
sources. In the case of CFUs, the lower number can
likely also be explained by loss of viability after
extended exposure to the atmospheric environment,
which must be considered a relatively hostile environ-
ment for bacteria in terms of risk of desiccation, low
temperatures, and high UV radiation.
The estimation of bacterial numbers also provides
information for the critical aspect of sampling bacteria
in an environment with low microbial density. In this
study, each sample was quality controlled by sterility
checking the sampling equipment and by enumerating
the bacterial cells in the buffer used to rinse the vortex
chamber before sampling. In the Swedish study, some
bacteria-shaped particles were detected in the micro-
scopic analyses of the controls, although the counts
were low and never linked to CFU detection on plates
or PCR product. Thus, we deduced this background as
caused by contamination of the material used to
prepare the microscopic slides. Interestingly, one of
the Danish air samples was counted at 0 CFU, which
suggests that our sampling method was reliable even
for an air mass whose bacterial density is below the
detection limit. In the crucial test of detection of DNA
Table 4 Backward trajectory analyses of the air masses sampled between 08:00 and 10:00 h local time along the sampling transect
by the Baltic Sea coast, Sweden, in summer 2009
Sampling
date
Source areas
0–12 h 12–36 h 36–96 h
No 1: Jul 14 Southern Sweden
Mainly forest
Denmark (?inner waters) and North Sea
Mainly agriculture and ocean
North Sea, southern England, and Atlantic Ocean
Mainly marine water and agriculture
No 2: Jul 16 Southern Sweden
Mainly forest
Denmark (?inner waters) and Germany
Mainly agriculture and ocean
Germany and France
Mainly agriculture and forest
No 3: Jul 22 Southern Sweden
Mainly forest
Denmark and North Sea
Mainly agriculture and ocean
Central England, Ireland, and Atlantic Ocean
Mainly agriculture, forests, and marine water
No 4: Aug 2 O
¨land and Baltic
Sea
Mainly water
Baltic Sea, Denmark (inner waters), and
Germany
Mainly ocean
North Sea, central England, Ireland, and Atlantic
Ocean
Mainly marine water
No 5: Aug 8 Baltic sea
Mainly water
Baltic countries
Mainly agriculture and forest
West Russia (from Baltic countries to Northern
parts)
Mainly forests, and tundra
No 6: Aug 9 Baltic sea
Mainly water
Baltic sea and Baltic countries
Mainly ocean, agriculture, and forest
Latvia, Russia, and Finland
Mainly forests
Fig. 3 Backward trajectories of the air masses sampled
between 08:00 and 10:00 h local time along the sampling
transect near O
¨land, Sweden, in summer 2009
c
492 Aerobiologia (2012) 28:481–498
123
Aerobiologia (2012) 28:481–498 493
123
and PCR products in the controls, the results were also
negative. Thus, we feel confident that the sampling
equipment was handled in a way that avoided sample
contamination. Although the total sampling volume
was as high as 100–400 m
3
of air, however, we were
still working near the detection limit, at least in the
case of CFU enumeration. We have previously tested
devices that sample approximately 2–20 m
3
of air
during a 2-h flight, and these sampling volumes
yielded no CFUs, that is, when using a modified
version of the BioCapture 650 impactor (ICx Tech-
nologies, Arlington, VA) (Fahlgren et al. 2011) or the
Airport MD8 (Sartorius, Goettingen, Germany). Con-
sistently, with the previous devices, we were also
unable to collect bacterial DNA in quantities sufficient
for sequence analysis.
The molecular analyses revealed that sequences
belonging to the genera Pseudomonas and Sphingo-
monas were present in great abundance in several of the
samples collected near the Swedish Baltic Sea coast.
These genera have previously been found to dominate
the boundary layer’s air mass (sampled 25 m above
ground) in the same geographic region (Fahlgren et al.
2010) and have also been found in air samples collected
approximately 1 m above the ground in Bergen, Nor-
way (Fahlgren et al. 2011), in dust collected in the
northern Caribbean (Griffin et al. 2003), in rural
environments in France (Maron et al. 2005), in Boulder,
CO (Fierer et al. 2008), in cloud water sampled on peaks
of the French Alps (Amato et al. 2007), and in airborne
dust from Africa deposited in the Pyrenees (Herva
`setal.
2009). Pseudomonas spp. and Sphingomonas spp. are
common in both aquatic and terrestrial environments, so
their presence in air, also above the ABL, is not
unexpected. However, the relative contributions of
these genera, up to 46%, are surprisingly high, and this
type of dominance is rarely found in possible source
communities (Janssen 2006;Pinhassietal.1997). The
ubiquitous and high presence of these genera may
therefore be indicative of a particular capacity to survive
in the atmospheric environment. It should also be noted
that identical Sphingomonas-like sequences were found
on several sampling dates, indicating that a single
species of this genus may occasionally dominate the
airborne community.
Many species and genera were found in low
abundance, but their presence in the atmosphere is
still of general interest. For example, Erwinia persicina
(Gonza
´lez and Rodicio 2007), Agrobacterium
tumefaciens, and Pantoea agglomerans, bacterial
species known as plant pathogens, were found in
several samples. Bacteria belonging to the genus
Clostridium were found in four samples, which is
somewhat surprising since this genus exclusively
includes species that are obligate anaerobes. The
finding of this genus in air possibly indicates the
presence of spores or persistent DNA belonging to
dead bacteria. Yet another interesting finding was the
presence of sequences most closely resembling Aur-
antimonas sp. in three of the samples. When analyzing
the sequences using the BLAST tool (Altschul et al.
1990), we found that an identical 16S rRNA gene
sequence (over the approximately 600-bp stretch
compared) has previously been isolated from clouds
sampled at Puy de Do
ˆme, France, at 1,500 m above sea
level (a.s.l.) (Amato et al. 2007). Since we sampled
above the ABL, this may indicate a more general
presence of this bacterium in layers that are relevant to
cloud formation. Bacterial species known for their
INA, for example, Pseudomonas syringae, as well as
Pseudomonas graminis, which has been found to carry
a gene homologue to known INPs, were also found on
several occasions (Nejad et al. 2006). It is unknown
whether the potential role of bacteria as condensation
nuclei or ice nuclei is determined by the stage of cloud
development at which they become a part of the cloud
particles, that is, during cloud initiation or precipitation
formation. If it is only during cloud initiation, then the
ABL bacteria may be a factor affecting weather
development in temperate regions (e.g., the region of
the present sampling area). If it is during precipitation
formation, the ABL bacteria may be a factor affecting
weather development in all regions of the world.
Therefore, the diversity and amount of airborne
bacteria above the ABL may be relevant to the amount
and distribution of precipitation worldwide. This,
however, is largely unexplored and is a subject for
future investigations.
The finding of numerous 16S rRNA genes from plant
chloroplasts and mitochondria on July 14 and August 9
is a distinct signal of a terrestrial source contribution in
the sampled air mass. However, when comparing the
back trajectories with the clone libraries, no simple
correlation is evident between the abundance of plant-
related sequences and the path of the air mass. This is
probably because, on all sampling dates, the sampled air
mass had passed over both sea and land areas over the
previous 4 days. It should be emphasized that trajectory
494 Aerobiologia (2012) 28:481–498
123
calculation is a conceptual tool and that the calculated
trajectories can only suggest possible source areas—not
determine the exact origin. In addition, single trajecto-
ries are known to be highly uncertain. This uncertainty
was reduced by applying the trajectory matrix method
(Skjøth et al. 2007), which in 2010 was adopted as a
standard methodology in the online version of the
HYSPLIT trajectory model (Draxler and Rolph 2010).
The trajectories therefore suggest possibly important
source regions for the sampled community. For exam-
ple, the clone library collected on July 16 deviated
markedly from those of other dates by containing a
higher fraction of gram-positive genera as well as
abundant cyanobacterial sequences. This sample had a
very distinct source profile, and the trajectories suggest
that this is the only sample for which the air masses had
traveled mainly over agricultural and forested areas of
central Europe (Fig. 3and Table 4).
When examining bacterial genera versus time, it
appears that variation in the airborne bacterial community
is not random. For example, Alphaproteobacteria,
primarily represented by Sphingomonas sp., appeared
continuously and in increasing numbers over the sam-
pling period (Fig. 2). This type of seasonal succession in
the microbial community is common in the marine
environment (Pinhassi and Hagstro
¨m2000) and in soils
(Lipson and Schmidt 2004). We foresee three possible
explanations for the apparent variation: (1) it reflects
annual succession in a source community; (2) it reflects a
succession of species in the airborne community; and (3)
the apparent succession is a result of random variation.
The first option can essentially be ruled out, since the back
trajectories indicate different sources for the different
sampling dates, and succession in the different source
communities is not likely correlated between habitats or
over geographical scales. Could the apparent variation
then be the result of the succession of species in the air,
that is, is it the result of bacterial growth in the
atmosphere? Air has traditionally been considered a
hostile environment in which microorganism survival
requires the return to land or sea. There are, however,
measurements of active bacterial growth in rainwater
(Sattler et al. 2001) as well as confirmation of bacterial
growth under laboratory conditions mimicking the cloud
environment (Vaı
¨tilingom et al. 2010). An observation
that adds plausibility to alternative two was the finding of
highly pigmented bacteria in the air samples; on average
84%, which is considerably higher than the proportion of
pigmented bacteria in, for example, surface soils or
waters or on ground vegetation (Du et al. 2006;Holmand
Jensen 1972). Non-fluorescent pigmentation, that is,
formation of colonies with the colors yellow, orange or
red, has been identified as a bacterial adaptation against
stress from UV radiation (Schlegel and Jannasch 1991;
Singer and Ames 1970). In this context, it should also be
noted that species belonging to the genus Sphingomonas,
found to dominate the 16S rRNA clone libraries in this
study, are known to be highly pigmented (Balkwill et al.
2006). Therefore, the high proportion of pigmented
colonies supports the notion that the airborne microbiota
could experience growth under selective pressure by UV
radiation. Thus, alternative two cannot be excluded,
although, with the limited number of samples presented
here, we also cannot rule out the third possibility, that is,
that the apparent variation is simply the result of random
events.
In conclusion, we have demonstrated that the air
above the ABL contains a highly variable 16S rRNA
gene diversity, including both bacteria- and plant-
related sequences, which has not been reported to date.
Several bacterial species and genera of interest for
cloud formation were detected. Notably, we found a
clear dominance of Sphingomonas spp. and Pseudo-
monas spp., indicating that these omnipresent species
may be particularly successful in the atmospheric
environment. Our findings document for the first time
the number and identity of bacteria above the ABL.
Their importance to convective and stratiform precip-
itation is, however, unknown, as the precise role of
biological particles remains under investigation.
Acknowledgments This study was funded by the European
Commission (PASR 2006), project AeroBactics (grant
agreement no. SEC6-PR-214400), and by the Swedish
Research Council for Environment, Agricultural Science and
Spatial Planning (FORMAS), grant no. 214-2008-1113. Niels
Bohse Hendriksen made valuable comments to this article. We
thank Lotte Frederiksen, Anne Grethe Holm-Jensen, Tina
Thane, Pia Petersen, Kilian Smith, and Tina S
ˇantl Temkiv for
their technical assistance. We wish to acknowledge our late
coauthor and colleague Runar Thyrhaug for his innovative idea
of using vacuum cleaner technology in the field of aerobiology.
References
Altschul, S. F., Gish, W., Miller, W., Myers, E. W., & Lipman,
D. J. (1990). Basic local alignment search tool. Journal of
Molecular Biology, 215(3), 403–410.
Aerobiologia (2012) 28:481–498 495
123
Amato, P., Parazols, M., Sancelme, M., Laj, P., Mailhot, G., &
Delort, A. M. (2007). Microorganisms isolated from the
water phase of tropospheric clouds at the puy de dome:
Major groups and growth abilities at low temperatures.
FEMS Microbiology Ecology, 59(2), 242–254.
Ashelford, K. E., Chuzhanova, N. A., Fry, J. C., Jones, A. J., &
Weightman, A. J. (2005). At least one in twenty 16 s rrna
sequence records currently held in public repositories
estimated to contain substantial anomalies. Applied and
Environmental Microbiology, 12, 7724–7736.
Balkwill, D. L., Fredrickson, J. K., & Romine, M. F. (2006).
Sphingomonas and related genera. In M. Dworkin, S. Fal-
kow, E. Rosenberg, K. H. Schleifer, & E. Stackebrandt
(Eds.), The prokaryotes: A handbook on the biology of
bacteria (pp. 605–630). Singapore: Springer-Science.
Bauer, H., Kasper-Giebl, A., Lo
¨flund, M., Giebl, H., Hitzen-
berger, R., Zibuschka, F., et al. (2002). The contribution of
bacteria and fungal spores to the organic carbon content of
cloud water, precipitation and aerosols. Atmospheric
Research, 64, 109–119.
Bowers, R. M., Lauber, C. L., Wiedinmyer, C., Hamady, M.,
Hallar, A. G., Fall, R., et al. (2009). Characterization of
airborne microbial communities at a high-elevation site
and their potential to act as atmospheric ice nuclei. Applied
and Environmental Microbiology, 75, 5121–5130.
Brandt, J., Christensen, J. H., Frohn, L. M., & Berkowicz, R.
(2001a). Operational air pollution forecasts from regional
scale to urban street scale. Part 1: System description.
Physics and Chemistry of the Earth Part B-Hydrology
Oceans and Atmosphere, 26, 781–786.
Brandt, J., Christensen, J. H., Frohn, L. M., & Berkowicz, R.
(2001b). Operational air pollution forecasts from regional
scale to urban street scale. Part 2: Performance evaluation
Physics and Chemistry of the Earth Part B-Hydrology
Oceans and Atmosphere, 26, 825–830.
Brodie, E. L., DeSantis, T. Z., Parker, J. P. M., Zubietta, I. X.,
Piceno, Y. M., & Andersen, G. L. (2007). Urban aerosols
harbor diverse and dynamic bacterial populations. Pro-
ceedings of the National Academy of Sciences, USA,
104(1), 299–304.
Cho, B. C., & Hwang, C. Y. (2011). Prokaryotic abundance and
16 s rrna gene sequences detected in marine aerosols on the
east sea (korea). FEMS MicrobiologyEcology, 76, 327–341.
Christner, B. C., Cai, R., Morris, C. E., McCarter, K. S., Fore-
man, C. M., Skidmore, M. L., et al. (2008a). Geographic,
seasonal, and precipitation chemistry influence on the
abundance and activity of biological ice nucleators in rain
and snow. Proceedings of the National Academy of Sci-
ences, USA, 105, 18854–18859.
Christner, B. C., Morris, C. E., Foreman, C. M., Cai, R., &
Sands, D. C. (2008b). Ubiquity of biological ice nucleators
in snowfall. Science, 319, 1214–1214.
Despres, V., Nowoisky, J., Klose, M., Conrad, R., Andreae, M.
O., & Po
¨schl, U. (2007). Molecular genetics and diversity
of primary biogenic aerosol particles in urban, rural, and
high alpine air. Biogeosciences Discussions, 4, 349–384.
Di Giorgio, C., Krempff, A., Guiraud, H., Binder, P., Tiret, C., &
Dumenil, G. (1996). Atmospheric pollution by airborne
microorganisms in the city of marseilles. Atmospheric
Environment, 30(1), 155–160.
Draxler, R. R., & Rolph, G. D. (2010). Hybrid single-particle
lagrangian integrated trajectory (hysplit). Model access
via noaa arl ready website: http://ready.Arl.Noaa.Gov/
hysplit.Php.
Du, H. L., Jiao, N. Z., Hu, Y. H., & Zeng, Y. H. (2006). Diversity
and distribution of pigmented heterotrophic bacteria in
marine environments. FEMS Microbiology Ecology, 57,
92–105.
Fahlgren, C., Bratbak, G., Sandaa, R.-A., Thyrhaug, R., &
Zweifel, U. L. (2011). Diversity of airborne bacteria in
samples collected using different devices for aerosol col-
lection. Aerobiologia, 27, 107–120.
Fahlgren, C., Hagstro
¨m, A
˚., Nilsson, D., & Zweifel, U. L.
(2010). Annual variations in the diversity, viability, and
origin of airborne bacteria. Applied and Environmental
Microbiology, 76, 3015–3025.
Fang, Z., Ouyang, Z., Zheng, H., Wang, X., & Hu, L. (2007).
Culturable airborne bacteria in outdoor environments in
beijing, china. Microbial Ecology, 54(3), 487–496.
Fierer, N., Liu, Z. Z., Rodriguez-Hernandez, M., Knight, R.,
Henn, M., & Hernandez, M. T. (2008). Short-term tem-
poral variability in airborne bacterial and fungal popula-
tions. Applied and Environmental Microbiology, 74(1),
200–207.
Franc, G. D., & DeMott, P. J. (1998). Cloud activation charac-
teristics of airborne erwinia carotovora cells. Journal of
Applied Meteorology, 37, 1293–1300.
Gonza
´lez, A. J., & Rodicio, R. (2007). Erwinia persicina
causing chlorosis and necrotic spots in leaves and tendrils
of pisum sativum in southeastern spain. Plant Disease, 91,
460.
Griffin, D. W., Kellogg, C. A., Garrison, V. H., Lisle, J. T.,
Borden, T. C., & Shinn, E. A. (2003). Atmospheric
microbiology in the northern Caribbean during African
dust events. Aerobiologia, 19, 143–157.
Hertel, O., Skjøth, C. A., Brandt, J., Christensen, J. H., Frohn, L.
M., & Frydendall, J. (2003). Operational mapping of
atmospheric nitrogen deposition to the Baltic Sea. Atmo-
spheric Chemistry and Physics, 3, 2083–2099.
Hertel, O., Skjøth, C. A., & Fenger, J. (2009). Atmospheric
physics, chapter 9. In J. Fenger, & J. C. Tjell (Eds.), Air
pollution—From a local to a global perspective (p. 488).
Polyteknisk forlag.
Herva
`s, A., Camarero, L., Reche, I., & Casamayor, E. O. (2009).
Viability and potential for immigration of airborne bacteria
from Africa that reach high mountain lakes in Europe.
Environmental Microbiology, 11, 1612–1623.
Herva
`s, A., & Casamayor, E. O. (2009). High similarity between
bacterioneuston and airborne bacterial community com-
positions in a high mountain lake area. FEMS Microbiol-
ogy Ecology, 67, 219–228.
Holm, E., & Jensen, V. (1972). Aerobic chemoorganotrophic
bacteria of a Danish beech forest. OIKOS, 23, 248–260.
Huber, T., Faulkner, G., & Hugenholtz, P. (2004). Bellerophon:
A program to detect chimeric sequences in multiple
sequence alignments. Bioinformatics, 20, 2317–2319.
Janjic, Z. I. (1990). The step-mountain coordinate: Physical
package. Monthly Weather Review, 118, 1429–1443.
Janjic, Z. I. (1994). The step-mountain eta coordinate model:
Further developments of the convection, viscous sublayer,
496 Aerobiologia (2012) 28:481–498
123
and turbulence closure schemes. Monthly Weather Review,
122, 927–945.
Janssen, P. H. (2006). Identifying the dominant soil bacterial
taxa in libraries of 16 s rrna and 16 s rrna genes. Applied
and Environmental Microbiology, 72, 1719–1728.
Kaarakainen, P., Meklin, T., Rintala, H., Hyva
¨rinen, A.,
Ka
¨rkka
¨inen, P., Vepsa
¨la
¨inen, A., et al. (2008). Seasonal
variation in airborne microbial concentrations and diver-
sity at landfill, urban and rural sites. CLEAN - Soil, Air,
Water, 36(7), 556–563.
Kasprzyk, I., Myszkowska, D., Grewling, L., Stach, A., Sa
´iko-
parija, B., Skjøth, C. A., et al. (2011). The occurrence of
ambrosia pollen in rzeszow, krakow and poznan poland:
Investigation of trends and possible transport of ambrosia
pollen from ukraine. International Journal of Biometeo-
rology, 55, 633–644.
Kawahara, H. (2002). The structures and functions of ice crys-
tal-controlling proteins from bacteria. Journal of Biosci-
ence and Bioengineering, 94, 492–496.
Kellogg, C. A., & Griffin, D. W. (2006). Aerobiology and the
global transport of desert dust. Trends in Ecology & Evo-
lution, 21(11), 638–644.
Kobashigawa, Y., Nishimiya, Y., Miura, K., Ohgiya, S., Miura,
A., & Tsuda, S. (2005). A part of ice nucleation protein
exhibits the ice-building ability. FEBS Letters, 579,
1493–1497.
Lighthart, B. (1997). The ecology of bacteria in the alfresco
atmosphere. FEMS Microbiology Ecology, 23, 263–274.
Lipson, D. A., & Schmidt, S. K. (2004). Seasonal changes in an
alpine soil bacterial community in the Colorado Rocky
mountains. Applied and Environmental Microbiology, 70,
2867–2879.
Ludwig, W., Strunk, O., Westram, R., Richer, L., Meier, H.,
Kumar, Y., et al. (2004). Arb: A software environment for
sequence data. Nucleic Acids Research, 32, 1363–1371.
Mancinelli, R. L., & Shulls, W. A. (1978). Airborne bacteria in
an urban-environment. Applied and Environmental
Microbiology, 35(6), 1095–1101.
Maron, P. A., Lejon, D. P. H., Carvalho, E., Bizet, K., Leman-
ceau, P., Ranjard, L., et al. (2005). Assessing genetic
structure and diversity of airborne bacterial communities
by DNA fingerprinting and 16 s rdna clone library. Atmo-
spheric Environment, 39(20), 3687–3695.
Martin, G. M., Johnson, D. W., Jonas, P. R., Rogers, D. P.,
Brooks, I. M., & Barlow, R. W. (1997). Effects of airmass
type on the interaction between warm stratocumulus and
underlying cumulus clouds in the marine boundary-layer.
The Quarterly Journal of the Royal Meteorological Soci-
ety, 123, 849–882.
Mo
¨hler, O., DeMott, P. J., Vali, G., & Levin, Z. (2007).
Microbiology and atmospheric processes: The role of
biological particles in cloud physics. Biogeosciences, 4,
1059–1071.
Muryoi, N., Matsukawa, K., Yamade, K., Kawahara, H., &
Obata, H. (2003). Purification and properties of an ice-
nucleating protein from an ice-nucleating bacterium,
pantoea ananatis kuin-3. Journal of Bioscience and Bio-
engineering, 95, 157–163.
Nejad, P., Ramstedt, M., Granhall, U., Roos, S., & McIvor, I.
(2006). Biochemical characterization and identification of
ice-nucleation-active (ina) willow pathogens by means of
biolog microplate, ina gene primers and pcr-based 16 s
rrna-gene analyses. Journal of Plant Diseases and Pro-
tection, 113, 97–106.
Pinhassi, J., & Hagstro
¨m, A
˚. (2000). Seasonal succession in
marine bacterioplankton. Aquatic Microbial Ecology, 21,
245–256.
Pinhassi, J., Zweifel, U. L., & Hagstro
¨m, A
˚. (1997). Dominant
marine bacterioplankton species found among colony-
forming bacteria. Applied and Environmental Microbiol-
ogy, 63, 3359–3366.
Porter, K. G., & Feig, Y. S. (1980). The use of dapi for identi-
fying and counting aquatic microflora. Limnology and
Oceanography, 25(5), 943–948.
Pruppacher, H. R., & Klett, J. D. (1997). Microphysics of clouds
and precipitation (2nd ed.). Dordrecht: Kluwer.
Reche, I., Ortega-Retuerta, E., Romera, O., Pulido-Villena, E.,
Morales-Baquero, R., & Casamayor, E. O. (2009). Effect
of Saharan dust inputs on bacterial activity and community
composition in mediterranean lakes and reservoirs. Lim-
nology and Oceanography, 54, 869–879.
Sattler, B., Puxbaum, H., & Psenner, R. (2001). Bacterial growth
in supercooled cloud droplets. Geophysical Research Let-
ters, 28, 239–242.
Schlegel, H. G., & Jannasch, H. W. (1991). Prokaryotes and
their habitats. In A. Balows, K. H. Schleifer, H. G. Truper,
& M. Dworkin (Eds.), The prokaryotes: A handbook on the
biology of bacteria: Ecophysiology, isolation, identifica-
tion, applications (pp. 90–91). New York: Springer.
Seinfeld, J. H., & Pandis, S. N. (1998). Atmospheric chemistry
and physics: Chap. 14 New York, USA: Wiley.
Shaffer, B. T., & Lighthart, B. (1996). Survey of culturable air-
borne bacteria at four diverse locations in Oregon: Urban,
rural, forest and coastal. Microbial Ecology, 34, 167–177.
Sikoparija, B., Smith, M., Skjøth, C. A., Radisic, P., Milkovska,
S., Simic, S., et al. (2009). The Pannonian plain as a source
of ambrosia pollen in the balkans. International Journal of
Biometeorology, 53, 263–272.
Simu, K., Holmfeldt, K., Zweifel, U. L., & Hagstrom, A
˚. (2005).
Culturability and coexistence of colony-forming and sin-
gle-cell marine bacterioplankton. Applied and Environ-
mental Microbiology, 71(8), 4793–4800.
Singer, C. E., & Ames, B. N. (1970). Sunlight ultraviolet and
bacterial DNA base ratios. Science, 170, 822–826.
Skjøth, C. A., Hertel, O., & Ellermann, T. (2002). Use of the
acdep trajectory model in the Danish nation-wide back-
ground monitoring programme. Physics and Chemistry of
the Earth, 27, 1469–1477.
Skjøth, C. A., Sommer, J., Brandt, J., Hvidberg, M., Geels, C.,
Hansen, K., et al. (2008). Copenhagen—A significant
source of birch (betula) pollen? International Journal of
Biometeorology, 52, 453–462.
Skjøth, C. A., Sommer, J., Stach, A., Smith, M., & Brandt, J.
(2007). The long range transport of birch (betula) pollen
from Poland and Germany causes significant pre-season
concentrations in Denmark. Clinical and Experimental
Allergy, 37, 1204–1212.
Smith, M., Skjøth, C. A., Myszkowska, D., Uruska, A., Mal-
gorzata, P., Stach, A., et al. (2008). Long-range transport of
ambrosia pollen to Poland agricultural and forest meteo-
rology. Agricultural and Forest Meteorology, 148,
1402–1411.
Aerobiologia (2012) 28:481–498 497
123
Soltis, P. S., Soltis, D. E., & Doyle, J. J. (1998). Molecular
systematics of plants ii: DNA sequencing. Boston: Kluwer.
Stach, A., Smith, M., Skjøth, C. A., & Brandt, J. (2007).
Examining ambrosia pollen episodes at poznan
˜(Poland)
using back-trajectory analysis. International Journal of
Biometeorology, 51, 275–286.
Staley, J. T., & Konopka, A. (1985). Measurement of in situ
activates of non photosynthetic microorganisms in aquatic
and terrestrial habitats. Annual Review of Microbiology,
24, 647–660.
Sun, J., & Ariya, P. A. (2006). Atmospheric organic and bio-
aerosols as cloud condensation nuclei (CCN): A review.
Atmospheric Environment, 40, 795–820.
Vaı
¨tilingom, M., Amato, P., Sancelme, M., Laj, P., Leriche, M.,
& Delort, A. M. (2010). Contribution of microbial activity
to carbon chemistry in clouds. Applied and Environmental
Microbiology, 76, 23–29.
Zhou, J., Swietlicki, E., Hansson, H. C., & Artaxo, P. (2002).
Submicrometer aerosol particle size distribution and
hygroscopic growth measured in the amazon rain forest
during the wet season. Journal Of Geophysical Research,
107, 10 pp.
ZoBell, C. E. (1946). Marine microbiology: A monograph on
hydrobacteriology. Waltham, MA: Cronica Botanica.
498 Aerobiologia (2012) 28:481–498
123
