• Atmosphere;
  • cloud water;
  • microorganisms;
  • fungi;
  • cold environment


  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References

This work constitutes the first large report on aerobic cultivable microorganisms present in cloud water. Seven cloud-event samples were collected at the Puy de Dôme summit, and cultivation was performed leading to the isolation of 71 bacterial, 42 fungal and 15 yeast strains. Most of the fungi isolated were of Cladosporium or Trametes affiliation, and yeasts were of Cryptococcus affiliation. Bacteria, identified on the basis of their 16S rRNA gene sequence, were found to belong to Actinobacteria, Firmicutes, Proteobacteria (Alpha, Beta and Gamma subclasses) and Bacteroidetes phyla, and mainly to the genera Pseudomonas, Sphingomonas, Staphylococcus, Streptomyces, and Arthrobacter. These strains appear to be closely related to some bacteria described from cold environments, water (sea and freshwater), soil or vegetation. Comparison of the distribution of Gram-negative vs. Gram-positive bacteria shows that the number of Gram-negative bacteria is greater in summer than in winter. Finally, a very important result of this study concerns the ability of half of the tested strains to grow at low temperatures (5°C): most of these are Gram-negative bacteria, and a few are even shown to be psychrophiles. On the whole, these results give a good picture of the microbial content of cloud water in terms of classification, and suggest that a large proportion of bacteria present in clouds have the capacity to be metabolically active there. This is of special interest with respect to the potential role of these microorganisms in atmospheric chemistry.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References

Clouds can be defined as atmospheric air masses in which water is condensed around particles in solid (ice crystals) or liquid form. They are low-temperature ‘aquatic’ environments involved in the large-scale transport of matter and energy, and so participate in the aerial connection between ecosystems, although it is still questionable whether clouds themselves can be considered as ecosystems.

It is assumed that the liquid and super-cooled water from which the free tropospheric clouds are composed provides a better temporary habitat for living airborne cells than dry air, where desiccation can be a limiting factor for growth. Thus cloud droplets may provide a medium in which these cells can divide, as suggested by Dimmick et al. (1979), Fuzzi et al. (1997) and Sattler et al. (2001). However, cloud water presents some specific characteristics such as acidic pH (generally from 3 to 7), high oxidative capacity, the presence of toxic compounds such as formaldehyde, high light (including UV) exposure, and relatively low temperatures (from −15 to 10°C at Puy de Dôme – see As a consequence, cells may require special physiological properties to remain alive in such an environment. To date, very few data are available concerning the microbial population in clouds. The reported total number of microorganisms in cloud water ranges from about 103 to 105 cells mL−1 of cloud water (Sattler et al., 2001; Bauer et al., 2002; Amato et al., 2005). Fuzzi et al. (1997) searched for cultivable cells in fog water at low altitude, and retrieved only three bacterial genera (Pseudomonas, Bacillus and Acinetobacter), and several fungi and yeasts. In tropospheric clouds, we previously observed (Amato et al., 2005) a more diversified cultivable population, consisting of many bacterial and fungal phyla. Many bacterial colonies were pigmented or spore-forming, and, as shown by their identification, some were closely related to strains isolated from cold environments. This cultivable fraction was found to represent only a small proportion of the total cells (<1%), but our measurements of ATP concentration in cloud water support the conclusion that most of the bacteria were still alive and certainly metabolically active (unpublished data). In addition, Amato et al. (2005) have shown that these bacterial strains contain the enzymatic equipment necessary to transform monoacid compounds (acetate, lactate, formate) as well as formaldehyde and methanol, which are present in relatively large concentrations in cloud water or are of primary interest in atmospheric chemistry (Suzuki et al., 1998; Marinoni et al., 2004). Our results, together with those reported by Ariya and coworkers (Ariya et al., 2002; Ariya & Amyot, 2004) in the case of diacid compounds, suggest a possible role of microorganisms in atmospheric chemistry. In order to confirm this hypothesis further, more data are needed describing the microbial population in clouds. We therefore present here the identification of culturable bacteria, yeasts and fungi isolated from seven cloud-water samples collected during different periods of the year. Growth ability at low temperatures was also investigated. This work constitutes the first large description of microbial living content to be found in cloud water.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References

Cloud sampling

The method and site of sampling are identical to those described in Amato et al. (2005). Sampling took place at the Puy de Dôme summit (1464 m a.s.l.), which is frequently covered by clouds, using a sterilized cloud-droplet impactor (Kruisz et al., 1993). The Puy de Dôme is fairly isolated from the other mountains of the Dôme mountain range, and the summit is mostly covered by grass. We therefore do not expect any interaction prior to sampling between cloud droplets and vegetation. In addition, the absence of dust particles in cloud water indicates a limited contribution of local soils to the cloud composition.

Bacteria, yeast and fungi isolation

Volumes of 0.1 mL−1 of cloud water were plated directly after sampling onto three different nutrient agar media, namely Trypcase Soy (TS; Biomerieux, Marcy l'Etoile, France), used as a general medium, R2A (Reasoner & Geldreich, 1985; DIFCO, Le Pont de Claix, France) for oligotrophic microbial strain recovery, and Sabouraud (DIFCO, Le Pont de Claix, France), a medium suited to fungal growth. Triplicates were performed for each medium at both 27°C and 15°C. After 3–5 days of dark aerobic incubation, each morphologically distinct colony was isolated by transplantation on the same media to obtain pure cultures.

A total of 128 microbial strains were isolated from seven cloud events, sampled between December 2003 and September 2004, including 71 bacterial, 42 fungal and 15 yeast strains.

Growth temperature tests for bacteria

The potentialities of strains isolated as colonies to grow at 5, 17 and 27°C were investigated. First, each bacterial strain was incubated at 5, 17 and 27°C in one of its specific media (R2, Emerson or Trypcase Soy broths); and second, 1 mL−1 was transferred into 20 mL−1 of the same medium and incubated once again at 5, 17 and 27°C. Therefore each strain was tested under different temperatures for preculture and for subsequent culture. Growth was monitored by measuring OD at 575 nm for about 120 h (5 days), and the slopes obtained during the exponential phases were used to calculate growth rates. In the case of filamentous bacteria and fungi these measurements were not possible as these organisms formed macroscopic pellets of cells during growth.

Strain identifications

Fungi and yeasts were isolated from single colonies and their identifications were carried out using microscopic observations and physiological tests at the Centraalbureau voor Schimmelcultures (CBS, Utrecht, the Netherlands). Bacteria were discriminated from eukaryotic cells (yeasts) by rapid microscopic observations and identified according to their 16S rRNA gene sequence. Cell pellets obtained after centrifugation of liquid pure cultures were resuspended in phosphate-buffered saline (PBS) solution, and their total genomic DNA was extracted using an Easy DNA Kit (Invitrogen sarl, Cergy Pontoise, France). Extracts were checked by gel electrophoresis, and 16S RNA genes were amplified by PCR. This step was carried out using universal primers for Eubacteria: F8-Eub (5′-AGA GTT TGA TCM TGG CTC-3′) and 1492r-Univ (5′-GNT ACC TTG TTA CGA CTT-3′) (Humayoun et al., 2003), in which M corresponds to A or C, and N to one of the four nucleotides indifferently. About 100 ng of genomic DNA and 1.5 U of Taq Polymerase (QBiogene, Illkirch, France) were used. PCR was performed as follows: 25 cycles of 30 s at 94°C for DNA denaturizing, 30 s at 55°C for hybridization with primers, and 90 s at 72°C for elongation, preceded by 5 min at 94°C and completed by 7 min at 72°C. PCR products were checked by gel electrophoresis, and purified on column using a Strataprep Purification Kit (Stratagene, Amsterdam, the Netherlands). They were finally freeze-dried and sequenced by capillary electrophoresis (MWG-Biotech, Roissy CDG, France), using the previously described primer F8-Eub. Comparisons of sequences with those included in GenBank were performed with the blastn interface, available at, in order to obtain the closest neighbours and gain information about the isolation source of these neighbours.

Results and discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References

Description of cultivable fungi and yeasts

Table 1 gives detailed identification of the isolated yeast and fungal strains and their occurrence in our samples. Among the 42 fungi and the 15 yeasts isolated, 13 and three, respectively, were taxonomically identified. Although culturable methods present well-known limitations, a remarkable point is the wide variety of the retrieved strains, belonging to Ascomycota, Basidiomycota and Deuteromycota groups. Amongst Ascomycota, genera Cladosporium (four strains), Aspergillus (two strains, including a teleomorph form Eurotium), and Penicillium (two strains) appear to be the most abundant. In the Basidiomycota phylum, seven Polyporales (genera Ganoderma, Phanerochaeta, Polyporus and four strains of Trametes), six Agaricales (one from the genus Schizophyllum and five strains not precisely identified) and one Boletales (Coniophora sp.) were found. Moreover, four strains belonging to the Deuteromycota embranchment were successfully identified (one Trichoderma sp. and three yeast strains: two Cryptococcus and one Pseudozyma).

Table 1.   Isolated fungal and yeast strains, and respective main isolation sources and characteristics
StrainPhylumCommon isolation sources and main characteristicsNumber of isolates
  1. ND, Not determined.

Acremonium sp.AscomycotaSoil, plant debris; saprophytic1
Aspergillus fumigatusAscomycotaAir, soil, decaying organic matter; saprophytic1
Botrytis cinereaAscomycotaPlants; phytopathogenic1
Cladosporium spp.AscomycotaWide spread, soil, plants; saprophytic4
Coniophora sp.BasidiomycotaWood; saprophytic1
Eurotium rubrumAscomycotaSoil; saprophytic; Teleomorph form of Aspergillus rubrobrunneus1
Fusarium avenaceumAscomycotaPlants; phytopathogenic1
Ganoderma sp.BasidiomycotaTrunks; intense sporing; parasitic and saprophytic1
Penicillium cf. sclerotiorumAscomycotaSoil; saprophytic1
Penicillium thomiiAscomycotaWide spread; sea water, plants, soil; saprophytic1
Phanerochaeta sp.BasidiomycotaWood; saprophytic1
Plectosphaerella cucumerinaAscomycotaSoil, decaying vegetation; saprophytic; teleomorph from of Plectosporium tabacinum1
Podospora sp.AscomycotaHerbivore dung; coprophilous1
Polyporus tuberasterBasidiomycotaDecaying wood; saprophytic1
Schizophyllum sp.BasidiomycotaDead wood; saprophytic1
Trametes spp.BasidiomycotaStumps, decaying wood; saprophytic4
Trichoderma citrinovirideDeuteromycotaDecaying wood and vegetation; saprophytic1
Verticillium nigrescensAscomycotaPlant; phytopathogenic1
Agaricales indet.Basidiomycota5
 Total 42
Cryptococcus albidusDeuteromycotaSoil, water, leaves, air; saprophytic1
Cryptococcus laurentiiDeuteromycotaWater, plants; saprophytic1
Pseudozyma sp.DeuteromycotaPlant materials; parasitic1
ND 12
 Total 15

All the cited genera are commonly found in natural environments, in soil and/or on vegetation. Most of them are saprophytic, growing on and causing decay of dead organic matter (plant and tree debris), or are phytopathogenic (Botrytis cinerea, Fusarium avenaceum, Verticillium nigrescens and Ganoderma sp.).

Fungal spores are frequently found in the air, and most of the genera cited here have previously been described in such environments. Genera such as Botrytis, Penicillium, Cladosporium and Aspergillus have been retrieved from indoor dry-air samples (Lin et al., 1999; Radon et al., 2002). Aspergillus and Cladosporium are certainly the genera most described in microbial studies of the atmosphere: in addition to their being detected indoors, they are widespread in outdoor air (Fulton, 1966b; Baxter & Cookson, 1983; Katial et al., 1997; Durand et al., 2002; Wittmaack et al., 2005), and also in nonprecipitating atmospheric water in the case of Aspergillus (Fuzzi et al., 1997; Amato et al., 2005). Concerning Cladosporium spp., not yet described in cloud water, their presence in our samples is not surprising in view of the number of studies reporting their presence in the air. Because in situ contamination can be excluded, it appears that, as for the other identified genera, Cladosporium spp. are integrated into droplets during cloud formation, meaning that such fungal spores can reach quite high altitudes despite their size, and can be transported over long distances. This remark also holds true for Trichoderma and Fusarium genera, collected at high altitude by Fulton (1966a), and also found in our cloud samples. Concerning yeasts, Cryptococcus members have been retrieved in alpine habitats, where they appeared as cold-adapted microorganisms (Bergauer et al., 2005).

Description of cultivable bacterial population

Table 2 presents a detailed view of bacterial genera cultured from each sample. The majority of bacterial isolates (61 from the 71 isolates) were successfully identified by sequencing their 16S rRNA gene and making a comparison with sequences in GenBank. Partial sequences have been lodged in GenBank, and accession numbers are specified in Tables 3 and 4.

Table 2.   Detailed bacterial strain isolation for each cloud event sampled
PhylumGenus12/12/ 200301/15/ 200401/21/ 200402/16/ 200406/24/ 200407/08/ 200409/23/ 2004Number of isolated strains
  • *

    Genus not accessible with sequenced fragment length.

ActinobacteriaStreptomyces 1 11 25
Arthrobacter12 1   4
Micrococcus1  11  3
Saccharothrix  1    1
Kocuria   1   1
Agromyces1      1
Cellulomonas1      1
Tetrasphaera1      1
Leucobacter1      1
Luteococcus 1     1
Agrococcus   1   1
Curtobacterium     1 1
Frigoribacterium      11
Nocardioides   1   1
Uncertain genus*1   1  2
FirmicutesStaphylococcus241    7
Bacillus11 1   3
Paenibacillus   1   1
Unidentified Bacilliales*    1  1
BacteroidetesPedobacter  1    2
Flavobacterium      11
Sphingobacterium      11
AlphaproteobacteriaSphingomonas sp.2  2  26
Methylobacterium   1   1
Aurantimonas   1   1
BetaproteobacteriaMassilia1      1
Zoogloea1      1
GammaproteobacteriaPseudomonas1 1 1238
Moraxella 2     2
Pantoa    1  1
Not determined 45 1   10
Table 3.   Gram-positive isolates
PhylumStrainSeq. length (bp)A.NDate of samplingClosest related neighbourA.N% homology (bp)Isolation sourceGrowth rate (OD/h)
at 5°Cat 17°Cat 27°C
  1. Genera and species names refer to the closest neighbours established through a blast search. Growth rates observed at 5°C, 17°C and 27°C are indicated where established. []Growth rates (OD/h during the exponential phase of growth) are −,<0.02; +, 0.02–0.049; ++, 0.05–0.29; +++, 0.3–0.59; ++++,>0.6.

  2. A.N, accession number in GenBank.

Actinobacteria3b-2869DQ51273812/12/2003Micrococcus luteus isolate CV39AJ71736899% (869/872)Alkaline groundwater   
3b-4860DQ51273912/12/2003Actinobacterium RG-9AY56157599% (852/860)High-Level Nuclear Waste-Contaminated   
3b-5679DQ51274012/12/2003Agromyces neolithicus strain 23-23AY50712898% (642/652)Not available   
3b-7836DQ51274212/12/2003Arthrobacter rhombi strain F98.3HR69Y1588599% (828/836)Greenland halibut   
3b-8510DQ51274312/12/2003Cellulomonas sp. 73NP9AB24268199% (507/510)Tomato leaf   
3b-18790DQ51275412/12/2003Tetrasphaera sp. Ellin150AF40899298% (777/792)Pasture soil   
3b-20848DQ51274812/12/2003Leucobacter aridicollis type strain L9AJ78104799% (842/847)Chromium contaminated environment   
5b-4698DQ51275501/15/2004Streptomyces albidoflavusAJ00209099% (698/699)Not available   
5b-7652DQ51275601/15/2004Luteococcus sanguinus strain CCUG 33897TAJ41675896% (599/619)Human blood   
5b-11566DQ51275801/15/2004Arthrobacter sp. isolate An29AJ551167100% (566/566)Deep sea sediment ++++++++
5b-15545DQ51276001/15/2004Arthrobacter sp. 255-8aAY44485299% (542/545)Ancient Siberian permafrost   
6b-3805DQ51276501/21/2004Saccharothrix tangerinus strain:MK27-91F2AB02003198% (793/806)Not available   
7b-1779DQ51276702/16/2004Streptomyces ciscaucasicus strain DSM 40275AY50851299% (774/781)Not available   
7b-2766DQ51276802/16/2004Agrococcus jenensisX92492100% (766/766)Frozen compost soil++
7b-4605DQ51276902/16/2004Micrococcus luteus CV31AJ71736799% (903/905)Alkaline groundwater   
7b-7659DQ51277202/16/2004Rhizosphere soil bacterium isolate RSI-25 (Nocardioides sp.)AJ25259299% (658/659)Rhizosphere soil 
7b-9647DQ51277302/16/2004Arthrobacter oxydansX8340899% (646/648)Not available ++
7b-12529DQ51277502/16/2004Kocuria rhizophilaY1626499% (528/529)Cattail of swamp plant   
12b-4398DQ51277906/24/2004Uncultured bacterium clone C13_D15AY99110598% (395/399)Mouse caecum ++++++
12b-6777DQ51278106/24/2004Micrococcus luteus CV31AJ717367100% (777/777)Alkaline groundwater ++++++
12b-10994DQ51278406/24/2004Streptomyces albidoflavusAJ00209099% (991/994)Not available   
13b-4934DQ51278707/08/2004Curtobacterium flaccumfaciens pv. BeticolaAY273208100% (934/934)Sugar beet+++
14b-9470DQ51279309/23/2004Streptomyces sp. N0143AY754724100% (470/470)Coastal sediment   
14b-12982DQ51279509/23/2004Streptomyces sp. N0130AY75472299% (980/984)Coastal sediment   
14b-13427DQ51279609/23/2004Frigoribacterium sp. GIC6AY43926299% (418/422)Glacial ice core from Greenland+++
Firmicutes3b-61393 DQ51274112/12/2003Bacillus simplex strain LMG 21002AJ316308100% (1397/1397)Mural painting environment++++++
3b-15953 DQ51275212/12/2003Staphylococcus epidermidis isolate CV64AJ71737799% (950/952)Alkaline groundwater   
3b-16942 DQ51275312/12/2003Bacterium G24 (staphylococcus sp.)AY34539599% (936/941)Green Lake water   
5b-1897 DQ51274901/15/2004Bacillus pumilus strain DF20AY46220599% (896/898)Not available 
5b-3896 DQ51275001/15/2004Bacterium G24 (staphylococcus sp.)AY345395100% (896/896)Green Lake water ++++
5b-6890DQ51275101/15/2004Staphylococcus cohnii isolate CV38AJ717378100% (889/889)Alkaline groundwater +++++++
5b-8898DQ51275701/15/2004Uncultured Staphylococcus sp. clone: MgMjD-019AB23451499% (897/898)Gut of a fungus-growing termite++++++
5b-16722DQ51276101/15/2004Staphylococcus equorum strain JH6DQ232735100% (722/722)Poultry farm++++++++
6b-1961DQ51276301/21/2004Bacterium G24 (staphylococcus sp.)AY34539599% (961/962)Green Lake water ++++++
7b-6667DQ51277102/16/2004Paenibacillus sp. isolate Eint3AM062705100% (667/667)Lichen+++
7b-11849DQ51277402/16/2004Bacillus licheniformis strain HK-1AY53096099% (847/850)Poultry waste ++
12b-7755DQ51278206/24/2004Bacillales bacterium s50-7-u8fDQ30531599% (750/755)Rice field soil+
Table 4.   Gram-negative isolates
PhylumStrainSeq. length (bp)A.NDate of samplingClosest related neighbourA.N% homology (bp)Isolation sourceGrowth rate (OD/h)
at 5°Cat 17°Cat 27°C
  1. Genera and species names refer to the closest neighbours established through a blast search. Growth rates observed at 5°C, 17°C and 27°C are indicated where established.

  2. Growth rates (OD/h during the exponential phase of growth) are −,<0.02; +, 0.02–0.049; ++, 0.05–0.29; +++, 0.3–0.59; ++++,.>0.6.

  3. A.N, accession number in GenBank.

Bacteroidetes6b-2804DQ51276401/21/2004Pedobacter sp. type strain HHS22AJ58342599% (792/800)Himalayan glacier   
14b-7898DQ51279109/23/2004Flavobacterium sp. EP101AF49364998% (862/873)River taff epilithon+
14b-8768DQ51279209/23/2004Uncultured Sphingobacterium sp. clone TM15_56DQ27936797% (733/748)Truffle++
Alphaproteobacteria3b-101007DQ51274412/12/2003Sphingomonas sp. Pmxh3DQ31473499% (1004/1007)Subnival plant   
3b-11609DQ51274512/12/2003Sphingomonas sp. TSBY-61DQ16616899% (608/609)Frozen soil   
7b-5682DQ51277002/16/2004Methylobacterium sp. strain: PB73AB220078100% (682/682)Terrestrial freshwater   
7b-13784DQ51277602/16/2004Sphingomonas sp. M3C203B-BAF39503199% (784/785)Antarctic subglacial lake Vostok+++++
7b-14703DQ51277702/16/2004Endophytic bacterium Enf15 (closely related to Aurantimonas sp.)DQ339602100% (703/703)Alpine subnival plants++++
7b-15858DQ51277802/16/2004Glacier bacterium FXS25 (Sphingomonas sp.)AY31516697% (841/859)New Zealand subglacial sediments++++++++
14b-5950DQ51278909/23/2004Sphingomonas sp. M3C203B-BAF39503199% (950/952)Antarctic subglacial lake Vostok   
14b-6968DQ51279009/23/2004Sphingomonas sp. isolate J05AJ86484299% (960/964)High mountain lake water+
Betaproteobacteria3b-121000DQ51274612/12/2003Zoogloea ramigera ATCC 25935X7491499% (921/930)Trickling filter of a sewage plant+++++++
3b-19869DQ51274712/12/2003Uncultured Massilia sp. clone KL-11-1-5AF40832698% (860/869)Clean-room facility   
Gammaproteobacteria3b-1858DQ51273712/12/2003Pseudomonas syringae pv. coryli strain NCPPB 4273AJ88984199% (852/854)Hazelnut   
5b-14815DQ51275901/15/2004Moraxella osloensisAY04337699% (814/815)Respiratory secretions   
5b-17808DQ51276201/15/2004Moraxella phenylpyruvicaAF00519299% (791/792)Spinal fluid   
6b-4813DQ51276601/21/2004Pseudomonas rhizosphaeraeAY15267398% (799/813)Rhizospheric soil of grasses++++
12b-5655DQ51278006/24/2004Pantoea agglomeransAY39501299% (652/656)Moth larval midgut ++++++
12b-8719DQ51278306/24/2004Pseudomonas syringae pv. atropurpureaAB00144099% (718/720)Graminaceous grass+++++++
13b-2887DQ51278507/08/2004Pseudomonas syringae pv. atropurpureaAB001440100% (887/887)Graminaceous grass   
13b-31004DQ51278607/08/2004Pseudomonas graminisY1115099% (1003/1005)Phyllosphere of grasses++++
14b-2898DQ51278809/23/2004Pseudomonas sp. SE22#1aAY26347799% (896/898)Alpine soil+++++
14b-10780DQ51279409/23/2004Pseudomonas sp. PH20AAY835585100% (780/780)Leaf spots of bean   
14b-14854DQ51279709/23/2004Pseudomonas viridiflava strain RMX23.1aAY574912100% (854/854)Arabidopsis thaliana++
General overview

Although only culturable bacteria were identified from the cloud-water samples, and not the total bacterial population, a large variety distributed among at least 28 genera can be described. Among the 61 isolates identified, 37 Gram-positive (61%) (25 Actinobacteria and 12 Firmicutes) and 24 Gram-negative (39%) (three Bacteroidetes, eight Alphaproteobacteria, two Betaproteobacteria and 11 Gammaproteobacteria) strains were found. The predominance of Gram-positive strains cultured from cloud water is consistent with previous works, showing that the Gram-positive cultivable fraction is generally more abundant in air than the Gram-negative one. Shaffer & Lighthart (1997) found about 80% of the cultivable population to be affiliated to Gram-positive phyla in dry air. This is supported by laboratory experiments showing that Gram-negative bacteria lose cultivability after being aerosolized (Heidelberg et al., 1997).

Actinobacteria was the most important phylum in terms of number of isolates and of genera represented: we found this affiliation for about half of the total number of genera identified. Genera Streptomyces (four strains), Arthrobacter (four strains) and Micrococcus (two strains) were the dominant ones, others (at least 11 different genera) being represented by only a single strain. The other phylum of Gram-positive bacteria (Firmicutes) was also highly represented, with respectively five and three distinct strains of the genera Staphylococcus and Bacillus.

Although the Gram-positive population is spread evenly amongst a great number of genera, the less-well represented Gram-negative group includes a few genera that are particularly dominant, namely Pseudomonas (Gammaproteobacteria) and Sphingomonas (Alphaproteobacteria), with respectively seven and five distinct strains isolated. These two genera are found to be the most abundant ones amongst all isolates. Only a few Betaproteobacteria and Bacteroidetes (also called Cytophaga–Flavobacterium–Bacteroides, CFB-group) were cultured from the samples, with respectively only two and three strains from these phyla.

Shaffer & Lighthart (1997) retrieved especially the genera Bacillus and Curtobacterium in their air samples. They also cited Arthrobacter and Micrococcus as being present. As in our observations of cloud droplets, Pseudomonas was the most abundant Gram-negative genus that they recovered from air samples. In another study focusing on airborne microbial content in an urban environment, cultivable Staphylococcus was found to be the predominant genus (Mancinelli & Shulls, 1978). Staphylococcus genus is also well represented in one of our samples, but back-trajectory plotting and chemical analysis (not shown) did not reveal any special urban influence on its composition. There is very little literature concerning the microbial content of cloud water; however, in fog, Bacillus and Pseudomonas seem to be the most commonly cultivated genera (Fuzzi et al., 1997). In a previous work based on only two cloud events at the Puy de Dôme (Amato et al., 2005), we also isolated Pseudomonas and Bacillus species, in addition to other genera such as Streptomyces, Micrococcus, Microbacterium, Flavobacterium, Kocuria, Arthrobacter, and Clavibacter.

Detailed identification of bacteria

Blast results obtained from the 16S rRNA gene sequences of our Gram-positive and Gram-negative isolates are respectively shown in Table 3 and 4. All sequences matched with known ones by a value at least as high as 97%, except for one that reached 96% (5b-7). Strains closely related to our isolates from cloud water were obviously from various origins, which can be regrouped as follows: aquatic environments (surface and ground water), extreme cold environments (polar and glacier samples, frozen soils), soil (sediment and other soils), vegetation (grass and plant leaves), and fauna (insects and other animals). A few strains cannot be included in these groups: strains 3b-4 (referenced as undetermined Actinobacterium) and 3b-20 (Leucobacter aridicollis) are Actinobacteria closely related to isolates from chromium- and nuclear-contaminated environments, while 3b-6 (Bacillus simplex) and 3b-19 (Massilia sp.) are neighbours to isolates recovered from mural painting and clean-rooms. In eight cases, the isolation source was not available.

The Gram-positive group appears to have a wide range of origins. Here we found highly diversified neighbour strains, from human-blood isolates (5b-7, Luteococcus sanguinus), to ancient Siberian permafrost (5b-15, Arthrobacter sp.) and sugar beet (13b-4, Curtobacterium flaccumfaciens). This suggests that, rather than being limited to carrying strains of environmental origin, clouds can also carry some potentially human pathogenic germs alive. Interestingly, many of the Gram-positive strains are close to isolates from zones such as deep-sea or coastal sediments, or groundwater. This particularly concerns the genera Micrococcus, Staphylococcus, Streptomyces and Arthrobacter. Genera such as Micrococcus, Bacillus and Staphylococcus are widespread in a large variety of environments, and are also present in cloud water.

Alphaproteobacteria isolates mainly belong to the genus Sphingomonas, but also to Methylobacterium (7b-5) and Aurantimonas (7b-14). All of them are closely related to microorganisms found in cold environments, and this group is broadly more often isolated from extreme cold environments (Christner et al., 2001; Zhang et al. 2002; Foght et al., 2004). Their physiological properties certainly make them fitted to living in an environment comparable to cloud water, as the low nutrient levels and cold conditions improve their survival rate. Gammaproteobacteria are for their part largely represented by Pseudomonas spp. (eight isolates out of 11), with the additional presence of two Moraxella spp. strains and one Pantoa agglomerans. Vegetation appears to be the major source of Pseudomonas, but this genus is known to be very ubiquitous and its presence was not surprising here. Furthermore, Pseudomonas have even been recovered from environments such as old glacial ice (about 750 000 years old) (Christner et al., 2003), testifying to their high resistance to extreme conditions. The strains from our cloud samples were often phytopathogenic, such as the example of P. syringae (3b-1, 12b-8 and 13b-2) (Sawada et al., 1997; Scortichini et al., 2002). For such strains, air constitutes an important way of dissemination, and their occurrence in our samples was to be expected. To retrieve P. syringae in cloud water is also of interest regarding the role that such a bacterium could play in the physics and chemistry of clouds. First, this species was found in Antarctic samples and is capable of cold adaptation (Seshu Kumar et al., 2002). Second, it has been demonstrated to have a very efficient ice-nuclei-forming activity (Cochet & Widehem, 2000), signifying its potential ability to initiate ice-crystal formation in supercooled clouds at relatively high temperatures. In the past, Schnell & Vali (1972) and Lindemann et al. (1982) suspected ice-nuclei to be produced in relation to vegetation, but P. syringae had not then been described in cloud samples. In addition, other species found in our samples are known to have a substantial ice-nuclei activity, for example bacteria of Pantoa affiliation and the fungi Fusarium avenaceum (Pouleur et al., 1992). We provide here proof of their regular presence in cloud droplets at different periods of the year, and it is therefore likely that such microorganisms are involved in microphysical processes occurring in clouds.

Other Gammaproteobacteria isolates (Moraxella) are more related to fauna emissions. The two last phyla of Gram-negative bacteria that were detected, Betaproteobacteria and Bacteroidetes, were more sporadic, but the presence of the Pedobacter sp. (6b-2), very similar to an isolate from a Himalayan glacier, can be noted (Shivaji et al., 2005).

Five bacterial strains were found several times in different samples. These were two Actinobacteria species [Streptomyces albidoflavus from a nonspecified source (5b-4 and 12b-10), and Micrococcus luteus CV31 from alkaline groundwater (7b-4 and 12b-6)], one Firmicutes [Bacterium G24, which is of Staphylococcus affiliation (3b-16, 5b-3 and 6b-1)], one Gammaproteobacteria [Sphingomonas sp. M3C203B-B from the Antarctic subglacial lake Vostok (7b-13 and 14b-5)], and one Gammaproteobacteria [Pseudomonas syringae pv. atropurpurea from graminaceous plant (12b-8 and 13b-2)].

Variations of bacterial composition with season

The results presented in Table 2 show that the total number of cultivable bacteria varies from one cloud event to another. In this section we briefly discuss the influence of the season on these variations. During summer, the proportion of Gram-positive cultivable bacteria (42.1% of the isolates) was less than that of Gram-negative ones (57.9% of the isolates). From summer to winter, the relative proportion of each group (Gram-positive and Gram-negative) changed, and the Gram-negative group became dominant.

In snow samples collected at an altitude of about 2500 m in Japan, Segawa et al. (2005) observed that the Gram-positive population contributed 71% of the whole bacterial flora in March, whereas it represented not more than 13% in June, and decreased to 8% in August. This is quite similar to what was observed in our cloud samples.

A first possible reason explaining the decrease of Gram-positive isolates in summer could be their greater sensitivity to the damaging effects of UV, more important in summer, as compared with Gram-negative ones. Recently, Aguoguéet al. (2005) showed that Gammaproteobacteria (Gram-negative) was the more represented group among the highly resistant to solar radiation bacteria isolated from the surface of the sea. A second reason could very possibly be the influence of summer on the preferential occurrence of microbial strains on vegetation which is far more developed in summer than in winter. Figure 1 gives a more detailed bacterial composition for the two periods, providing averages of the occurrence of each phylum per sample, and confirms our two previous hypotheses. First, it shows that the observed difference between winter and summer is in great part the result of a decrease in the Firmicutes proportion from 21% to 5%, combined with an increase in Gammaproteobacteria from 8% to 37%. Such a fact can be supported by the higher resistance of Gammaproteobacteria to solar radiation, as previously mentioned. Second, the development of vegetation in summer could also explain the increase of this phylum, which includes the Pseudomonas genus, well known to grow on plants.


Figure 1.  Global composition in aerobic cultivable bacteria during winter and summer. The contribution of each phylum to an averaged cloud-water sample is given with respect to the season.

Download figure to PowerPoint

Abilities to grow at low temperatures

Some isolated strains were tested for their aerobic growth at 5, 17 and 27°C. The results are presented in Tables 3 and 4. Atmospheric temperature is recorded continuously at puy de Dôme (VAISALA PT 100 sensor), showing an average value of 5.2±1.2°C. The lowest incubation temperature was therefore chosen in the range of this mean annual temperature. It is also interesting to note that winter-time temperatures at puy de Dôme regularly go down to −15°C. Twenty strains were tested for the three temperatures, and for eleven others only the 17 and 27°C tests could be performed because of technical problems. For convenience, we created four classes of growth ability according to the growth rates expressed as OD values per hour (see legends of Tables 3 and 4).

For two strains (7b-7 Nocardioides sp. and 5b-1 Bacillus pumilus), the growth rates remained very low whatever the temperature investigated.

Six strains out of 38 tested (about 16%) grew faster at 27°C than at 5 and 17°C. All of them are of Gram-positive affiliation: they include four Actinobacteria and two Firmicutes strains. Some strains are of special interest because they grow better at 17°C: 14b13 (Frigobacterium sp.), 5b8 (Staphylococcus sp.), 6b4 (Pseudomonas rhizosphaerae), and 3b12 (Zoogloea ramigera). The two last strains also grow at 5 and 27°C, and can thus be considered as psychrotolerant.

More interesting for this study of clouds are those strains able to grow at a low temperature close to the natural atmospheric environment, i.e. 5°C. This was the case for 11 strains out of 20 tested (about 55%), including five Gammaproteobacteria (Pseudomonas spp.), two Firmicutes (Paenibacillus sp., and Staphylococcus sp.), two Alphaproteobacteria (Sphingomonas spp.), one Betaproteobacteria (Zoogloea ramigera), and one Bacteroidetes (Flavobacterium sp.): the large majority are Gram-negative strains. It can be noted that three strains have a relatively high growth rate at 5°C: 3b12 (Zoogloea ramigera), 5b16 (Staphylococcus equorum), and 7b15 (Sphingomonas sp.). Finally, two strains are also remarkable because they grow at 5°C but not at 17 and 27°C, and can therefore be considered as true psychrophiles: 14b14 (Pseudomonas viridiflava) and 14b7 (Flavobacterium sp.), according to the definition of Morita (1975).

Clearly, Gram-negative strains are much more efficient at growing at low temperatures, because 75% (9/12) of those tested have this capacity. Among the Gram-positive strains, only 25% (2/8) presented this ability. This is consistent with observations made on iced environments such as the poles and high mountains, where the Gram-negative community is dominant (Alfreider et al., 1996; Brinkmeyer et al., 2003; Groudieva et al., 2004). These results are also comparable with those obtained by Groudieva et al. (2004), who found only a few truly psychrophilic strains among a population dominated by psychrotolerants in a polar environment. In addition, many colonies we isolated were yellow to orange pigmented (not shown), as is often reported for bacteria cultivated from cold environments (Fong et al., 2001; Foght et al., 2004).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References

This work is the first to report an extensive study of aerobic cultivable microorganisms present in atmospheric water. This water was collected at the top of the Puy de Dôme mountain, over a period of more than a year and a half corresponding to seven cloud events. Although culture methods do not allow all members of the population to be described, they give very valuable information about the physiology of microorganisms that are really alive in clouds. This is of special interest given that the cloud medium can prevent microorganisms from surviving because of the harsh conditions such as cold temperature, UV exposure, low pH, and oxidative environment.

General considerations can be drawn concerning this population: (i) most of the isolated fungi identified were of Cladosporium or Trametes affiliation, and yeasts were also present; (ii) bacteria were found to belong to Actinobacteria, Firmicutes, Proteobacteria (Alpha, Beta and Gamma subclasses) and Bacteroidetes phyla, with dominant occurrence of the genera Pseudomonas, Sphingomonas, Streptomyces, Staphylococcus and Arthrobacter. The population of aerobic cultivable bacteria is dominated by strains that are highly similar, regarding their 16S rRNA gene sequence, to some isolates from cold environments, water (sea and freshwater), soil and vegetation. This study demonstrates the great diversity of microbial population that can be found in clouds, even using only culture-dependent methods.

Comparison of the distribution of Gram-negative vs. Gram-positive bacteria as a function of season shows that summertime, with its higher light exposure, can favour the presence of Gram-negative strains, which are usually more resistant to UV damage. Finally, the numbers of the Gram-negative strains we isolated appear to be generally associated with vegetation that is obviously more present in summertime in Europe. All these parameters lead to a higher proportion of retrieved microorganisms related to Gram-negative strains during the summer.

A very important result of this study concerns the ability of an important proportion of the isolated bacteria to grow at temperatures compatible with those occurring in tropospheric clouds. These results suggest that a great proportion of bacteria present in clouds can be metabolically active in this medium. It is of special interest regarding the potential role of microorganisms in atmospheric chemistry. Work is in progress to test the abilities of these strains to transform organic compounds at low temperatures, and to determine the metabolic pathways involved under these conditions.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References

P.A. is the recipient of a scholarship from the French Ministry of Research. This work is supported by the National Program on Atmospheric Chemistry (PNCA) from the French CNRS and the ORE-BEAM program from the French Ministry of Research. The staff of the Microbiology Unit at the INRA of Clermont-Ferrand-Theix are acknowledged for their technical help in DNA isolation.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. References