Phylogenetic diversity of bacteria associated with Paleolithic paintings and surrounding rock walls in two Spanish caves (Llonín and La Garma)

Abstract

Bacterial diversity in caves is still rarely investigated using culture-independent techniques. In the present study, bacterial communities on Paleolithic paintings and surrounding rock walls in two Spanish caves (Llonín and La Garma) were analyzed, using 16S rDNA-based denaturing gradient gel electrophoresis community fingerprinting and phylogenetic analyses without prior cultivation. Results revealed complex bacterial communities consisting of a high number of novel 16S rDNA sequence types and indicated a high biodiversity of lithotrophic and heterotrophic bacteria. Identified bacteria were related to already cultured bacteria (39 clones) and to environmental 16S rDNA clones (46 clones). The nearest phylogenetic relatives were members of the Proteobacteria (41.1%), of the Acidobacterium division (16.5%), Actinobacteria (20%), Firmicutes (10.6%), of the Cytophaga/Flexibacter/Bacteroides division (5.9%), Nitrospira group (3.5%), green non-sulfur bacteria (1.2%), and candidate WS3 division (1.2%). Thirteen of these clones were most closely related to those obtained from the previous studies on Tito Bustillo Cave. The comparison of the present data with the data obtained previously from Altamira and Tito Bustillo Caves revealed similarities in the bacterial community components, especially in the high abundance of the Acidobacteria and Rhizobiaceae, and in the presence of bacteria related to ammonia and sulfur oxidizers.

1 Introduction

Caves are generally nutrient-poor biotopes with relatively stable and low temperatures and with high mineral concentrations. Therefore, caves can be considered extreme environments for life and provide ecological niches for highly specialized microorganisms. Hypogean environments are often severely resource-limited due to the absence of light, which prevents the growth of phototrophic microorganisms and plants. Several studies report that lithotrophic bacteria play an important role in some cave ecosystems by acting as primary producers and supporting growth of heterotrophic microorganisms, when light is completely absent. In these caves the ecosystem does not depend on organic input of epigean origin [1,2]. In some other caves, dripping water, visitors and animals can provide organic input, facilitating life of heterotrophic microorganisms [3–5]. Cultivation studies revealed the presence of bacteria, mainly affiliated with lithotrophic sulfur- and iron-oxidizing bacteria, nitrifying Proteobacteria, and also heterotrophic Proteobacteria, low G+C heterotrophic Gram-positive bacteria, and Actinomycetes [2–9]. Geomicrobiological activities in caves are no longer underestimated, since studies showed that bacterial metabolism leads to mineral precipitation processes, and/or dissolution of cave walls [10,11].

Furthermore, bacterial growth also affects painting pigments [12–15] and underlying rock material. Many caves such as the famous Spanish caves of Tito Bustillo and Altamira, decorated with Paleolithic paintings, are threatened by microbial colonization, which does not spare this valuable work of art. Information on microbial community structure in caves with Paleolithic paintings is rather limited, but necessary in order to gain more knowledge on its possible impact on the valuable paintings.

Culture-independent 16S rDNA sequence analyses opened the way to study bacterial communities in environmental samples without prior cultivation and revealed a significantly broader diversity of 16S rDNA sequence types than culture-based studies [16–18]. The combination of 16S rDNA clone libraries and community fingerprinting via the sequence-specific separation of polymerase chain reaction (PCR) amplicons by denaturing gradient gel electrophoresis (DGGE) has become a powerful tool for investigating natural bacterial communities [19]. Nevertheless, most studies of bacterial communities in caves have been based on cultivation techniques [4–7,10,12,20], while the 16S rDNA-based analysis of bacterial colonization in caves has been restricted to only a few studies [9,21–23]. The 16S rDNA approach opened the way to analyze bacterial colonization on medieval wall paintings and Paleolithic paintings without significantly affecting the paintings because only a small sample is required. The studies revealed diverse and unknown microbial colonization on the paintings [24–29]. Until recently, studies on microbial colonization in caves decorated with Paleolithic paintings were restricted to samples taken from the surrounding rock walls, while no samples taken from pigments could be investigated. In our previous studies [28,29] we had the opportunity to investigate bacterial colonization of painting pigments from Tito Bustillo and Altamira Caves applying culture-independent methods. The present study reports on bacterial colonization of the Paleolithic paintings from Llonín and La Garma Caves (northern Spain), where access is restricted to serve research purposes only. In both caves, wall rock surface areas near the Paleolithic paintings and some parts of the paintings show microbial colonization in the form of macroscopic spots consisting of colonies (1–2 mm). Bacterial communities of eight samples were analyzed using DGGE community fingerprinting and comparative phylogenetic 16S rDNA sequence analyses.

2 Materials and methods

2.1 Cave descriptions

Llonín Cave, located near the village of Llonín (Asturias, northern Spain), has been known for a long time. In 1957 the cave was adapted for storage and maturation of local cheese, which continued until 1970. In the following year the cave and the Paleolithic paintings were ‘officially’ discovered by a group of invited speleologists. Very limited archeological studies of the painting panels were carried out in 1971–1987. Between 1987 and 1997 the cave was visited once a year for archeological excavations, both inside and outside. The paintings (estimated age 22 000–10 000 BC) have been under study since 1997. In the lower part of the cave, several primitive abstract figures in red and black color, dating back to the Magdalenian period, are present. In the Llonín Cave artificial light has been installed for archeologists. The temperature in the cave at the sampling time (October 3, 1998) was 13°C and the relative humidity was 100%.

La Garma Cave was discovered in November 1995 under a hill near the village of Omoño (Cantabria, northern Spain) and is part of a complex karstic system of galleries situated at different heights and connected by vertical chasms. The cave indicates an extensive Paleolithic settlement with paintings from the initial Upper Paleolithic to the Magdalenian period and the ground is full of evidence of human activity (animal bones, charcoal fragments, etc.). The valuable paintings show horses, bison, deer, non-figurative traces and handprints. The cave is accessible only for archeological excavations. Between November 1997 and October 1998 the temperature in the hall of the Lower Gallery was quite constant at 13.2°C from February to August and at 13.3°C during the remaining months. By the time of the sampling (October 4, 1998) 64 persons had visited the cave since its discovery, with a total time of stay of 45 h and 30 min, a mean of three persons per visit and an average visit duration of 42 min. Relative humidity at the sampling place was 100%. La Garma Cave has no light, thus growth of phototrophic microorganisms is prevented.

2.2 Sampling and DNA extraction

Sample material was scraped off the rock or the paintings with a sterile scalpel under the supervision of the responsible archeologists. Samples C6–C8 were collected in Llonín Cave. Sample C6, which was taken from an area showing abstract paintings, consisted of a red paint layer material without apparent microbial colonization. Sample C7 was taken a few cm away from sample C6 and contained wall material without apparent microbial colonization. Sample C8 represented small, white, round macroscopic colonies. Bacterial colonies were sampled without affecting the supporting rock. Samples C9–C13 were collected in La Garma Cave. Sample C9 was a portion of a white round patch of a colony (about 3 cm diameter) taken from a black painting stroke of a non-identifiable quadruped situated close to the gallery. Sample C10 consisted of red paint layer material without apparent microbial colonization, and was taken from a painting area showing non-figurative traces. This area is located deep inside one of the galleries. Sample C11 represents a white-grayish macroscopic colony growing on a black rock surface a few cm from sample C9. Sample C12 was from yellow-grayish macroscopic colonies growing on a red rock surface. Sample C13 was a powdery white biolayer on a speleothem.

DNA was extracted from less than 1 mg subsamples as described previously for wall painting material [26]. Briefly, cell lysis was based on lysozyme and proteinase K digestion, sodium dodecyl sulfate treatment and freeze-thawing. Possible PCR inhibitors were complexed by hexadecyltrimethyl ammonium bromide and genomic DNA was purified by elution through silica gel membranes.

2.3 PCR amplification of 16S rDNA fragments, construction of 16S rDNA clone libraries and DGGE community fingerprinting

Three μl of the DNA extract was tested for PCR-amplifiable DNA with 16S rDNA-specific primers. Fragments corresponding to nucleotide positions 341–926 of the Escherichia coli 16S rDNA were amplified with the forward primer 341f and the reverse primer 907r as described previously [26]. Clone libraries were constructed by cloning 5 μl of the PCR product. Per sample, 100 recombinant clones were randomly screened for different 16S rDNA inserts using DGGE as described previously [26]. Inserts of clones, producing PCR products matching the most intense bands and some faint bands of the DGGE fingerprints, were sequenced.

For DGGE community fingerprinting, 200-bp fragments corresponding to nucleotide positions 341–534 (E. coli numbering) were amplified with the forward primer 341fGC and the reverse primer 518r [19]. PCR amplifications were carried out in 100 μl reactions with 5 μl PCR product of the first amplification as template DNA [26]. PCR products, obtained by nested PCR, were pooled (200 μl), precipitated with 96% ethanol, re-suspended in 15 μl sterile and double distilled H₂O and separated by DGGE. A reference marker containing 16S rDNA fragments of 10 known bacteria was loaded on the gel in order to allow gel-to-gel comparison. Gel electrophoresis was performed as described elsewhere [19] in a linear gradient of denaturant from 25% to 60% in a D GENE-System (Bio-Rad, Munich, Germany). After completion of electrophoresis, gels were stained in an ethidium bromide solution and documented with a UVP documentation system.

2.4 Sequencing of 16S rDNA inserts and phylogenetic analysis

For sequencing of clone inserts, 100 μl PCR product generated with vector-specific primers SP6 and T7 was purified with a QIAquick PCR Purification Kit (Qiagen, Hilden, Germany) and sequenced partially or completely with a LI-COR DNA sequencer Long Read 4200 [30]. Sequencing reactions were carried out by cycle sequencing with the SequiTherm™ system (Epicentre) with 2 pmol fluorescently labelled primer 341f and 5 U SequiTherm thermostable DNA polymerase.

16S rDNA sequences were compared with sequences of known bacteria listed in the EMBL nucleotide sequence database. The FASTA search option of the EMBL database was used to search for close evolutionary relatives [31]. The most similar sequences were used for multiple sequence alignments with ClustalW (http://www.ebi.ac.uk/clustalw/). Sequences of 16S rDNA clones derived from Tito Bustillo and Altamira Caves [28,29] were included in the analyses. Distance matrices and phylogenetic trees were calculated by the Jukes and Cantor [32] and neighbor-joining [33] algorithms using the programs DNADIST and NEIGHBOR, respectively, provided in the Phylogeny Inference Package PHYLIP 3.5 [34]. The stability of the groupings was checked by bootstrap analysis of 1000 replications. The tree was rooted using an archaeal sequence (Thermoproteus sp.) as the outgroup. Short sequences (200 bp) were excluded from the tree construction.

The ribosomal sequences obtained in this study have been deposited in the EMBL database under the accession numbers AJ421153–AJ421179 (Llonín Cave) and AJ421084–AJ421141 (La Garma Cave).

3 Results

3.1 PCR amplification and DGGE analyses

16S rDNA fragments were amplified and genetic fingerprinting of bacterial communities was performed using DGGE. Repeated electrophoresis analyses gave reproducible results (not shown). Figs. 1A and 2A show details of the DGGE community fingerprints of samples C6–C8 (Llonín Cave), C9–C13 (La Garma Cave), and of the reference marker (M). Figs. 1B and 2B show their schematic drawing. DGGE revealed complex bacterial communities with patterns containing about 11–28 visible DGGE bands per sample.

3.2 rDNA clone sequencing and phylogenetic analyses

Phylogenetic information on individual bacterial members present in the samples was obtained from ribosomal clone libraries (600 bp). The numbers of sequenced 16S rDNA clones per sample are summarized in Table 1 and the most closely related sequences listed in the EMBL database in Tables 2 and 3. Similarity values to 16S rDNA sequences listed in the EMBL database ranged from 83.5% to 100% to cultured bacteria (a total of 39 clones) and from 89.7% to 99.3% to environmental 16S rDNA clones (a total of 46 clones). Cloned sequence types affiliated with (i) the Proteobacteria, including members of the α, β, γ and δ subdivisions (41.1%), (ii) the Nitrospira (nitrite oxidizer) group (3.5%), (iii) the Cytophaga/Flexibacter/Bacteroides phylum (5.9%), (iv) green non-sulfur bacteria (1.2%), (v) the candidate division WS3 (1.2%), (vi) members of the Acidobacterium division (16.5%), (vii) Firmicutes (10.6%), and (viii) the Actinobacteria (20%), as summarized in Table 1. Phylogenetic trees including 600-bp ribosomal sequences recovered from the present study, those of their closest relatives from the EMBL database and ribosomal sequences obtained from previous studies in Altamira and Tito Bustillo Caves [28,29] are given in Figs. 3 and 4.

(i) Proteobacteria. Bacteria, affiliated with the Proteobacteria, were related to cultured members and to cloned sequence types of the α-Proteobacteria [Rhizobiaceae (number of clones, n=13), Sphingomonas group (n=3), and Rhodospirillaceae (n=2)], β-Proteobacteria [Comamonadaceae (n=1)], γ-Proteobacteria [Ectothiorhodospiraceae (n=5), Chromatiaceae (n=3), Nitrosococcus (n=2), Xanthomonas group (n=1), Pseudomonadaceae (n=1), Moraxellaceae (n=1), and Thermomonas (n=1)], and δ-Proteobacteria [Desulfobacterium (n=1), and Geobacter (n=1)] (Tables 2 and 3). In the Llonín Cave, members of the Rhizobiaceae and of the Sphingomonas group were the most abundantly identified Proteobacteria, while in La Garma Cave, Proteobacteria were represented mostly by the members of Rhizobiaceae, Ectothiorhodospiraceae and Chromatiaceae.

Most clones revealed more than 95% similarity to proteobacterial sequences. Three clones (C11-K7, C11-K33 and C8-K3) shared less than 90% similarity to known sequences. These clones seem to be representatives of novel taxa within the γ and δ subdivisions of the Proteobacteria, respectively. Some clones (C6-K21, C10-K25, C10-K39, C10-K48, C11-K11, C11-K16 and C13-K15) found in Llonín and La Garma Caves grouped with phylogenetically novel 16S rDNA sequence types recovered from previous, culture-independent studies in Tito Bustillo Cave [29] (Fig. 3, Tables 2 and 3). This group of bacteria seems to be highly abundant in La Garma and Tito Bustillo Caves.

(ii) Nitrospira group. Three clones recovered from Llonín and La Garma Caves were related to uncultured Green Bay ferromanganous micronodule bacteria with highest affiliation to Nitrospira. These clones were also related with more than 97% similarity to clones C3-K5 and C4-K18 from Tito Bustillo Cave (Fig. 3, Table 3).

(iii–v) Cytophaga/Flexibacter/Bacteroides phylum, green non-sulfur bacteria and candidate division WS3. In both caves, members of these phyla were rarely identified (seven clones) and showed low similarities (89.0–94.1%) to uncultured bacteria (Fig. 4A, Tables 1–3). Clone C10-K13 shared 96% similarity with soil clone PRR-10 [35]. PRR-10 is a member of a novel bacterial lineage WS3, recently discovered in a contaminated aquifer with no cultivated representatives thus far [36]. This lineage is described by a few 16S rDNA clones, recovered from an anoxic zone of a hydrocarbon- and chlorinated-solvent-contaminated aquifer (clone WCHA1-56) [36], from anoxic sediments in eastern Antarctica (clones ACE-39 and ACE-9) [37], and from anoxic bulk soil of flooded rice microcosms (PBS and PRR clones) [35]. Our finding of a member of the WS3 division in an aerobic environment is thus unexpected. However, sequence information of clone C10-K13 recovered from La Garma Cave is only 200 bp, and is thus not included in a phylogenetic tree.

(vi) Acidobacteria. Fourteen clones of La Garma Cave were affiliated with the Acidobacterium division. The Acidobacterium division was the second dominating phylogenetic group after the Proteobacteria identified in La Garma Cave. No Acidobacteria were found in Llonín Cave. Clones were most closely related to yet uncultured bacteria with similarities between 90.5 and 99.0% (Table 3). Concerning hypogean environments, Acidobacteria have been detected only in two other caves (Altamira and Tito Bustillo) [28,29]. Fig. 4B shows a phylogenetic tree of Acidobacteria recovered from La Garma, Altamira and Tito Bustillo Caves and their closest relatives. Ribosomal sequences grouped with acidobacterial subclusters a–c described by Ludwig et al. [38]. The described subclustering was a first attempt to show the intraphylogenetic structure of Acidobacteria. Most 600-bp sequences recovered from La Garma Cave grouped within subcluster c (n=8) and were most closely related to Tito Bustillo Cave clone C3-K13 (n=2), Roggenstein soil clone RB41 (n=4) and Arizona soils clone S023 (n=2).

(vii) Firmicutes. Nine clones were related to low G+C Gram-positive bacteria (Tables 2 and 3). Most of them were found in La Garma Cave (n=8). Six clones were related to Bacillus spp. with similarities between 96 and 100%. Three clones shared less than 90% similarity to Bacillus, Oscillospira and Clostridium. They most likely represent novel taxa within the low G+C Gram-positive bacteria. Fig. 3A shows a phylogenetic tree of 600-bp sequences.

(viii) Actinomycetes. Seventeen clones shared similarities between 89.4 and 99.6% to high G+C Gram-positive bacteria (Tables 2 and 3). Clones grouped with the members of the Actinosynnemataceae [Saccharothrix sp. (n=2)], Pseudonocardiaceae [Streptoalloteichus sp., Actinobispora sp., and uncultured bacterium related to Thermocrispum sp. (n=8)], Micrococcaceae [Arthrobacter sp. (n=1)], Streptomycetaceae [Streptomyces sp. (n=1)], Nocardioidaceae [Nocardioides sp. and uncultured bacterium related to Nocardioides sp. (n=2)], as well as Spirillospora sp. (n=1), and uncultured bacteria related to Micothrix sp. (n=1). Members of the Pseudonocardiaceae were the most abundantly recovered Actinobacteria. Clone C11-K18 was related with less than 90% similarity to Spirillispora sp. and most likely represents a novel phylogenetic taxon within the Actinobacteria. Fig. 4A shows a phylogenetic tree of 600-bp sequences, together with 16S rDNA sequences recovered from Altamira and Tito Bustillo Caves. Clones C6-K26, C6-K18, C7-K3, C7-K31, C12-K40 and C11-K10 formed a tight cluster that was most closely related to Streptoalloteichus sp. and Saccharothrix sp.

4 Discussion

While most microbiological investigations in caves are based on cultivation studies, the present report is a descriptive study on the bacterial colonization in two Spanish caves with a high cultural heritage interest due to their archeological remains and Paleolithic paintings. The culture-independent approach gave insight into the great, previously unexplored bacterial diversity in both caves. Investigated samples represented material without apparent microbial colonization from the Paleolithic paintings and surrounding walls as well as macroscopic colonies growing on the paintings and on the surrounding rock surfaces. The sampling sites on the paintings and rocks can be considered representative of either an iron-rich habitat (the red paintings based on iron oxides) or cave rocks. Sampling in the caves is essentially prohibited and so this study was a unique opportunity to gain insight into the microbial communities thriving in caves that are not subjected to mass tourism. Microbial communities were dominated by the members of Proteobacteria, followed by the Actinobacteria, Acidobacteria and Firmicutes. Members of the Cytophaga/Flexibacter/Bacteroides division, Nitrospira group, green non-sulfur bacteria, and candidate WS3 division were less abundant. Proteobacteria are commonly detected in caves [7,8,10]. The comparison of present results with the data of previous studies in Altamira and Tito Bustillo Caves [28,29] revealed high numbers of members of the Rhizobiaceae in all caves. The finding of bacteria grouping with the Nitrospira group and with bacteria related to putatively lithotrophic bacteria (sulfur and ammonia oxidizers) identified in La Garma Cave is similar to the results obtained from Tito Bustillo Cave [29] (Fig. 3). The presence of lithotrophic bacteria as a part of complex bacterial communities in caves has already been confirmed in some PCR-based studies and cultivation experiments [9,21–23,40]. Sulfur and sulfide oxidizers, iron and manganese oxidizers, sulfate reducers, and nitrifiers [9,11,21–23,39,40] appear abundant in caves. Sulfur-oxidizing bacteria play a role in the dissolution of limestone in caves with hydrogen sulfide-rich waters, contributing to cave enlargement. The extent to which bacteria contribute to the corrosion of limestone walls and to the enlargement of existing caves remains uncertain [11].

Our finding of bacteria related to the Acidobacterium division in La Garma Cave further confirmed the distribution of this monophyletic group in hypogean environments. The Acidobacterium division is a poorly studied phylogenetic division so far. It was defined by Ludwig et al. [38] on the basis of cloned 16S rDNA sequences from soil, freshwater sediments and activated sludge in many geographic locations. Its representatives are thought to be ecologically significant microorganisms in many ecosystems [38,41]. Their 16S rDNA sequences were phylogenetically related to the cultivated species Holophaga foetida, Geothrix fermentans, and Acidobacterium capsulatum. Acidobacteria have already been identified in two other Spanish caves (Altamira and Tito Bustillo) [28,29]. Like in these caves, the Acidobacteria were the second dominating group after the Proteobacteria in La Garma Cave, revealing that Acidobacteria contribute significantly to microbial colonization in these environments. It is unclear whether the anthropogenic impact due to the use of Llonín Cave as a cheese factory for 13 years had influenced the wall surface microflora and explains the absence of members of the Acidobacterium division.

Like in Altamira and Tito Bustillo Caves, the number of Gram-positive bacteria identified in Llonín and La Garma Caves was rather low – 30% of the 16S rDNA sequences were affiliated with Gram-positive bacteria related to the Bacillus/Clostridium group (low G+C Gram-positives) and to the Actinobacteria (high G+C Gram-positives). However, Gram-positive bacteria, and particularly actinomycetes, were the dominating isolates obtained with cultivation techniques [3–5]. In general, low G+C Gram-positive bacteria detected in caves were mainly represented by Bacillus, Clostridium, Kocuria, Microbacterium, Paenibacillus, and Staphylococcus[7,8,10,20]. In the present study, Bacillus was the most abundantly identified bacterium of this group. Actinomycetes were originally isolated from soils, but can be found in nearly every habitat. They also contribute a significant portion of isolated bacteria in hypogean environments [3–6,8,20].

Microbiological studies in caves decorated with Paleolithic paintings have so far been restricted to investigation of surrounding rock walls, whereas in this study, sample material was taken directly from Paleolithic paintings. Results revealed that complex bacterial colonization was not restricted only to visible colonies, but was present also in material without apparent microbial colonization. In view of the complex DGGE community patterns (Figs. 1 and 2), the limited number of 27 and 58 analyzed clones per cave described only a small portion of the phylogenetic diversity of the samples. Bacterial diversity shown in DGGE might be underestimated because of possible identical electrophoretic migration of sequences with multiple differences [42]. Strong bacterial colonization on valuable paintings is alarming since bacterial communities may have biocorrosive potential and cause irreversible damage. Microbial colonization and metabolic activities can lead to the formation of pigmented biofilms, to biomineralization, to the dissolution of metals by acids and chelating agents, to enzymatic reduction of metal oxides, and to the degradation and discoloration of painting pigments [13–15,43]. In general, iron oxide, charcoal, and black manganese earth were used as painting pigments for Paleolithic paintings. Among these, the red pigment iron oxide is especially endangered. Bacillus spp. and Arthrobacter viscosus, isolated from rock art paintings, were found to reduce hematite (iron oxide) in laboratory cultures [12]. In the present study, we also found relatives of Bacillus spp. and Acidobacteria in samples of black and red painting material, along with a high number of bacteria related to sulfur- and ammonia-oxidizing bacteria and Actinomycetes. Sulfur- and ammonia-oxidizing bacteria produce inorganic acids, which could also affect the painting painting pigments. Actinobacteria are also of special interest, as members of this group are known to destroy wall paintings by excretion of organic and inorganic metabolic products. However, the culture-independent approach applied in the present study allows only vague assumptions about the physiological properties of identified bacteria and the ecological and biodeteriorative role they might play.

The question arises if mass tourism in caves and the resulting changes of climatic conditions, the input of organic matter into the ecosystem and the growth of secondary colonizers brought in by visitors might change natural cave communities. It is known that natural communities in caves may become unbalanced when artificial light is introduced in the ecosystem, which promotes growth of phototrophic microorganisms. This is noticed in some parts of Tito Bustillo Cave, where most stalactites and stalagmites are covered by Cyanobacteria[10]. While Tito Bustillo and Altamira Caves have been visited extensively in the past (Altamira Cave received up to 3000 visitors per day in the 1970s, and Tito Bustillo Cave about 50 000 visitors per year), Llonín and La Garma Caves are not open to the public. Despite the different cave management, bacterial community composition in La Garma and in Tito Bustillo Caves does not differ significantly. It seems that bacteria are mainly intrinsic inhabitants with no or little anthropogenic impact on their presence. Nevertheless, management of caves that are open to human visitation must aim at maintaining the caves’ natural climatic conditions and eliminating artificial light, in order to preserve valuable Paleolithic paintings. It is likely that molecular sequence-based techniques will provide more information on diversity and structure of bacterial communities in cave environments and in future allow monitoring of changes in bacterial colony growth on Paleolithic paintings and surrounding walls.

Acknowledgements

This research was supported by the Austrian–Spanish Cooperation Agreement HU 1997-0035 and the MCyT project BTE2002-04492-C02-01. The facilities provided by the archeologists responsible for the caves are acknowledged.

References

Author notes