The vital activities of epilithic microorganisms contribute to undesirable changes in works of art. The organisms responsible for biodeterioration build complex structured communities, called biofilms, detracting from the aesthetic impression and eventually inducing chemical and physical damage . Qualitative and quantitative characterisation of biodeterioration demands an understanding of the substratum, organisms and abiotic factors implicated . Such information may help administrators to choose the appropriate preventive and eradication methods  for cultural patrimony at risk .
Roman catacombs are hypogean monuments that form part of a tourist route. Visited areas suffer intense changes caused by the visitors themselves and/or by the conditioning of the galleries for visits , the latter of which implies the use of artificial illumination. We ultimately aimed to produce innovative, non-destructive technologies to control and prevent uncontrolled growth of photosynthetic biofilms on rock surfaces.
Light quality is a determining factor in photosynthetic biofilm development, as light is fundamental to the metabolism and development of phototrophic microorganisms . Owing to the artificial illumination installed for visitors, hypogean areas earmarked for tourism suffer from more phototrophic activity – and hence, more biofilm growth – than non-tourist sights. Spectral confocal microscopy equipped with a lambda scan function has been used for 3D localisation and in vivo discrimination by fluorescence signatures of individual organisms from different phylogenetic groups (e.g. chlorophyta, bacillario-phyta, and phycoerythrin- and non-phycoerythrin-containing cyanobacteria) in complex communities .
In this research work, we also evaluated the effects of monochromatic green light (GL), nearly unused by photosynthetic organisms, on the adaptation and viability of phototrophic organisms, the structure of biofilms, and ultimately, on its efficacy in preventing biofilm growth in hypogean environments that are normally illuminated with artificial white light (WL). We chose monochromatic green light because it provides maximum sensitivity for human vision, thereby preserving the details and tonal values of cultural attractions.
Biofilms and lighting system
Natural biofilm samples were obtained from artificially illuminated surfaces in hypogean monuments (St. Callistus and Domitilla catacombs, Rome, Italy). The samples were maintained on a 2 mm layer of BG11 medium  at a nutrient concentration of 10 % and solidified with 1 % agar (Merck). Artificial biofilms were constructed by inoculating sterilised calcareous slabs, deposited on the aforementioned medium, with 1 g of a mixture of axenic cultures of Gloeothece membranacea (Rabenhorst) Bornet CCAP1430/3 (Pasteur Culture Collection, Paris, France) and green alga Chlorella sorokiniana Shih. & Krauss from the Culture Collection of Instituto Isla Cartuja (CSIC) (Sevilla, Spain). The slabs were placed in Petri dishes and kept at 19–22 °C under continuous green light (GL) (Narva LT 18 W/017 green TT, Narva, Czech republic) or white light (WL) (Chiyoda F 15 S daylight, Chiyoda Corporation, Japan) at a constant photon flux density of 20 µmol · m–2 · s–1 for 60 days. The emission spectrum of each lamp was measured with a LICOR Li-1800 (Lincoln, NE, USA) spectroradiometer.
Visualisation of biofilm structure
Confocal Scanning Laser Microscopy (CSLM) was performed with a Leica TCS SP2 in fluorescence and reflection modes. The reflection mode (excitation at 488 nm, and emission at 480 to 490 nm) allowed recording of reflective signals originating from inorganic solid material. Autofluorescence from photosynthetic pigments was viewed in the red channel using the 543 and 633 nm lines of an Ar/HeNe laser in the emission range of 590 to 800 nm. Extracellular polysaccharides (EPS) were labeled with the lectin Concanavalin-A-Alexa Fluor 488 (Molecular Probes, Inc., Eugene, OR, USA) (0.8 mM final concentration) and observed in the green channel (excitation at 488 nm line, and emission at 490 to 530 nm). Nucleic acids were specifically stained with the DNA-selective dye Hoechst 33258 (Sigma-Aldrich, St. Louis, MO, USA) and viewed in the blue channel (excitation at 351 and 364 nm and emission at 400 to 480 nm).
We acquired optical sections in both x-y planes obtained at different intervals (z step) along the optical axis. Images were presented as multichannel and 3D projections using Imaris software (Bitplane, Zürich, Switzerland). The projections were used to describe the spatial distribution of organisms within the biofilms as well as differences in biofilm thickness and architecture.
Pigment fluorescence analysis: lambda scans
Single cell pigment identification was performed according to Roldán et al. (2004b) . Wavelength scans were performed using the 488 nm line of an Ar laser. Each image sequence (wavelength scan) was obtained by scanning the same x-y optical section using a bandwidth of 20 nm for the emission. The x-y-l data set was acquired at the z position having maximum fluorescence. Gains and offsets were the same for each field and remained constant throughout the scanning process.
The variation in intensity of a particular spectral component was represented on the screen using a false-colour scale (warm colours represent maximum intensities, whereas cool colours show low intensities). Mean Fluorescence Intensity (MFI) of the x-y-l data sets was measured using Leica Confocal Software, version 2.0. The region of interest (ROI) function of the software was used to determine the spectral signature of a selected area from the scanned image. For the fluorescence analysis, ROIs of 1 µm2 taken from the thylakoid region inside the cell were set in each x-y-l stack of images. The mean and standard errors for all the ROIs were calculated.
Biofilms from Roman catacombs
In naturally or artificially illuminated areas, biofilm-forming phototrophic microorganisms developed profusely on surfaces composed of plaster, fresco, tufa, brick or mortar (Figure 1a, 1b). The biofilms (Figure 2) primarily comprised coccal and filamentous cyanobacteria and mosses, but also included diatoms, fungi hyphae and actinobacteria . The biofilms were generally porous and very heterogeneous in thickness (from 80 to 550 µm), as determined by the three dimensional projections from autofluorescence and EPS (Figure 2). For the majority of biofilms, the EPS were most abundant at the upper layer (Figure 3a). Interestingly, the reaction of morphologically similar phototrophic cells to Con-A varied with their relative position inside the biofilm.
Living phototrophic and the heterotrophic microorganisms were distinguished by coupling the results from Hoechst 33258 DNA staining with those from pigment autofluorescence studies. The nucleic acid stain did not show an extended heterotrophic bacterial community. However, these communities were highly developed in zones having weak photosynthetic pigment fluorescence. The reflection channel revealed the presence and thickness of inorganic hard substrates, such as calcified sheaths (Figure 2c); coupled to autofluorescence it discriminates empty calcified sheaths
Changes in the structure and composition of arti-ficial biofilms submitted to either GL or WL were evaluated using three-dimensional images, whereby photosynthetic pigments (PP) fluoresce red, and polysaccharide-bound Con-A fluoresces green (Figure 4). Both GL and WL biofilms were stratified (Figure 4).
GL: Slabs maintained under GL contained clusters of two to four cells of Gloeothece membranacea surrounded by a relatively thick sheath, only labeled for Con-A to the oldest external layers (Figure 4a). Vacuolised thylakoids of G. membranacea were seen as empty spaces inside the cells. C. sorokiniana was nearly absent, and the few remaining cells had weak pigment fluorescence.
WL: Slabs kept under WL displayed clusters of G. membranacea mixed with sub-spherical cells of C. sorokiniana. Both organisms had higher respective pigment fluorescence than in the case of GL. There were dense clusters of 8 to 16 cells of G. membranacea, which were sometimes further aggregated and surrounded by an apparently compact sheath that was strongly labeled (Figure 4b). Biofilms grown in WL (Figure 4b) were thinner (PP thickness = 3.7 ± 1.53 µm) than those grown in GL (PP thickness = 4.65 ± 0.9 µm), and included a compact bottom layer of small, irregularly dense colonies of G. membranacea while C. sorokiniana grew close to the surface.
Fig. 4: Confocal projections of cyanobacterial biofilms grown under GL and WL. Colour key: PP, red; EPS, green; NA, blue. a. Gloeothece membranacea colonies grown in GL. The sheath was not compact, and only the external layer was labeled by Con-A lectin (arrow). b. G. membranacea colonies grown in WL. The cells were grouped in dense clusters of up to sixteen cells and surrounded by a compact sheath (arrowhead). Chlorella sorokiniana was found mostly in the upper layer. T= Biofilm surface. Scale bar = 10 µm.
Biofilms from Roman catacombs
Extended focus images of stratified biofilms showed the differential distribution in depth of the biofilm microorganisms (Fig. 5). For each biofilm, the corresponding emission spectra at 488 nm excitation wavelength are shown on the right (Figure 5b).
Biofilm BF1: Thin filaments of Leptolyngya sp. were horizontally oriented on top of wide Scytonema julianum (Figure 5a). Both cyanobacteria had a broad emission peak at 658.4 ± 3 nm from the overlap of Chlorophyll a and phycobiliproteins, phycocyanin (PC) and allophycocyanin (APC) (Figure 5b). Additionally, Leptolyngybya sp., but not S. julianum, presented an emission peak (579.7 ± 3.8 nm) attributable to the presence of phycoerythrin (PE) (Figure 5b). We did not observe changes in the emission peak of particular specimens from the same taxa (e.g. Scytonema julianum when covered by thick sheaths, EPS or calcareous investments).
Biofilm BF2: CSLM revealed two layers. Diadesmis gallica – Bacillariophyta – was mainly concentrated on the top of the biofilm (Figure 5a). Their highest maximum, at 676.2 ± 5 nm (Figure 5b), did not coincide with the emission peak of the other groups due to the presence of Chlorophyll c.
The unidentified Chroo-coccal formed a discontinuous bottom layer (Fig. 5a) and presented the same spectral shape as Leptolyngybya sp. (BF1). The cyanobacteria presented higher mean fluorescence intensity (MFI) in the range from 640 to 740 nm than did Bacillario-phyta (Figure 5b). The cyanobacteria also presented high MFI at 577 to 580 nm (Figure 5), due to PE.
Fig. 5: CSLM projections and fluorescence properties of two aerophytic biofilms from Roman catacombs. a. Each image represents the autofluorescence emitted in the range 590–775 nm (Excitation wavelength = 543 nm). BF1. Bio-film formed by Scytonema julianum and Leptolyngbya sp. BF2. Stratified biofilm, consisting of two strata, the upper epilithic layer composed of colonies of Diadesmis gallica, and a lower layer formed by Chroococcal colonies. b. Spectral profile from single cells of each species from a lambda scan (excitation wavelength = 488 nm, steps = 50). The differences among the emission profiles of the biofilms indicate the presence of different groups of algae and cyanobacteria. T= Biofilm surface. Scale bar = 10 µm (published in: Applied Environmental Microbiology 70 (2004) 3745–3750).
MFI and the half-band width of the spectra from both species, at 488 nm excitation wavelength, were different for GL than for WL whereas the spectrum shape was identical (Figure 6).
Gloeothece membranacea: For the GL sample, the highest maximum corresponded to Chl a (ca. 670.7 nm), whereas the emission bands for WL samples were considerably higher (659.3 to 666.4 nm), corresponding to the phycobiliproteins PC and APC (Figure 6b). In both GL and WL, G. membranacea showed PE fluorescence at an emission wavelength of ca. 580 nm when excited at 488 nm. In GL, the peak position closely matches the main photosystem II (PSII) emission peak at 685 nm, with a secondary maximum near 730 nm that corresponds to photosystem I (PSI) emission bands (Figure 6b). Significant differences between the MFI of GL and WL treated samples were found for the emission range of 573.6 to 590.7 nm, in which PE emits, and 653.6 to 659.3 nm, in which PC and APC emit. We can assume that GL influences fluorescence properties by decreasing the net rate of fluorescence.
Chlorella sorokiniana: At 488 nm excitation wavelength, significant differences in the Chl a maximum were not observed for samples of either treatment (Figure 7).
Fig. 6: Fluorescence properties from Gloeothece membranacea single cells after illumination with different light qualities (GL and WL). a. Optical sections taken from a lambda scan (excitation wavelength = 488 nm, steps = 50) corresponding to the emission peak of Chl a and phycobiliproteins (PE, phycoery-thrin; PC, phycocyanin; APC, allophycocyanin). b. Spectral profile representing the mean fluorescence intensity spectra for G. membranacea under GL and under WL. Scale bar = 10 µm (published in: Applied Environmental Microbiology 72 (2006) 3026–3031).
Fig. 7: Fluorescence properties from Chlorella sorokiniana single cells after illumination with different light qualities (GL and WL). a. Optical sections taken from a lambda scan (excitation wavelength = 488 nm, steps = 50) corresponding to the emission peak of chlorophylls. b. Spectral profile representing the mean fluorescence intensity spectra for Chlorella sorokiniana under GL and under WL. Note the decrease in mean fluorescence intensity of the pigments from GL treatment relative to the WL control. Scale bar = 10 µm (published in: Applied Environmental Microbiology 72 (2006) 3026–3031).
Organisms and biofilm structure
CSLM is an excellent technique for the study of the distribution of microorganisms and their biodeteriorative effects . It has been employed for imaging phototrophic biofilms and mats in various habitats, catacombs  and other dimly lit aerophytic environments such as sinkholes and caves . For biofilm samples, the use of multiple excitation and detection wavelengths at different focal depths enabled us to three-dimensionally visualise target-specific elements such as molecules (e.g. DNA and pigments), structures (e.g. surfaces, matrices, sheaths and filaments) and properties (e.g. stage of cell division, growth and senescence).
To date, CSLM with the lambda-scan function has only been used to determine the optimal detection and separation of emission spectra for either new or known fluorochromes. In this study, we used CSLM with lambda scan to create a powerful tool with a broad scope of applications. Fluorescence detection allows the description of a complex community in terms of physiological state  and quantification of biomass . Chlorophyll a is a photosynthetic pigment found in most plants, algae and cyanobacteria. In addition, most cyanobacteria use phycobiliproteins (phycoerythrin, phycocyanin and allophycocyanin) to capture light energy. Using lambda scan function, species belonging to one phylogenetic group showed spectrophotometric profiles distinguishable in shape and pigment emission peak from other phylogenetic groups . The main features of the technique are: (i) analysis for both global and single fluorescent pixels, providing their three dimensional localisation in vivo; (ii) direct analysis of fluorescent pigments from a single cell in situ in thick samples without isolation; (iii) establishing a simultaneous relationship between fluorescence properties, morphology and position inside complex microbial assemblies; and (iv) discrimination of cells with particular fluorescence signatures within the colony, and correlation with individual cell states.
Methods for preventing biodeterioration
The elimination of microbial colonisation from works of art is not trivial. Most of the treatments to remove phototrophic biofilms consist of biocide applications which often have undesirable effects. Hence there is a pressing need for alternative, low-risk methods . We think that the key to controlling biodeterioration is the use of measures that prevent favourable growth conditions for microflora. Our results suggest that the use of monochromatic GL, which is poorly absorbed by the majority of photosynthetic organisms, is useful for preventing green sickness in hypogean environments. Only some cyanobacteria are able to harvest sufficient light energy to survive in extremely dim ambient light or in GL  although they do not thrive . This indicates that illuminating hypogean environments with GL instead of WL can reduce the diversity of moss and algae species.
There is a need to further develop techniques to monitor the natural conditions in which phototrophic biofilms grow, and to evaluate the qualitative and quantitative changes that preventive or eradicative treatments have on biofilms. CSLM is advantageous for use in protecting cultural patrimony sites because it allows the in vivo exploration of very small samples (e.g. samples from precious works of art), and provides information in real time and space. This method combined with a new technique such as white light laser or fluorescence lifetime imaging microscopy (FLIM)  creates the potential for a broad range of new experiments regarding photosynthetic microorganisms. Moreover, spectral CSLM has enabled us to identify the pigments used by photosynthetic microorganisms at specific wavelengths [15, 17]. We are currently exploiting this information to develop illumination alternatives for threatened sites of cultural patrimony.
We thank Dr. Oliva for his contributions to statistical calculations. This work was partially supported by the EU Energy, Environment and Sustainable Development Program within the framework of the CATS project, contract EVK4-CT-2000-00028.
- Hoffmann L: Caves and other low-light environments: Aerophytic photoautotrophic microorganisms, 835–843. In: Bitton G (ed), Encyclopedia of Environmental Microbiology. John Wiley & Sons, New York (2002).
- Roldán M, Clavero E, Castel S, Hernández-Mariné M: Biofilms fluorescence and image analysis in hypogean monuments research. Arch Hydrobiol, Algol Stud 111 (2004) 127–143.
- Albertano, P, Moscone D, Palleschi G, Hermosin B, Saiz-Jiménez C, Sánchez-Moral S, Hernández-Mariné M, Urzí C, Groth I, Schroeckh V, Gallon JR, Graziottin F, Bisconti F, Giuliani F: Cyanobacteria attack rocks (CATS): Control and preventive strategies to avoid damage caused by cyanobacteria and associated microorganisms in Roman hypogean monuments, 151–162. In: Saiz-Jiménez C (ed), Molecular Biology and Cultural Heritage. Swets & Zeitlinger, Lisse (2003).
- Agarossi G, Ferrari R, Monte M: Microbial deterioration in the hypogea: the subterranean neo-Pythagorean Basilica of Porta Maggiore in Rome. In: Felix G (ed), Proceedings of the 5th International Congress on Deterioration and Conservation of Stone. Lausanne: Presses Polytechniques Romandes, 597–605 (1985).
- Albertano P.: Deterioration of Roman hypogea by epilithic cyanobacteria and microalgae, 1303–1308. In: Guarino A (ed), Science and Technology for the Safeguard of Cultural Heritage in the Mediterranean Basin 2 (1998). CNR Luxograph, Palermo.
- Ariño X, Hernández-Mariné M, Saiz-Jiménez C: Colonization of Roman tombs by calcifying cyanobacteria. Phycologia 36 (1997) 366–373.
- Ciferri O.: Microbial degradation of paintings. Appl. Environ. Microbiol 65 (1999) 879–85.
- Kumar R, Kumar AV: Biodeterioration of stone in tropical environments. Los Angeles: Paul Getty Trust (1999).
- Roldán M, Clavero E, Hernández-Mariné M: Aerophytic biofilms in dim habitats. In: Saiz-Jiménez C (ed), Molecular Biology and Cultural Heritage, Swets & Zeitlinger, Lisse (2003).
- Caldwell DE, Korber DR, Lawrence JR: Imaging of bacterial cells by fluorescence exclusion using scanning confocal laser microscopy. J Microbiol Methods 15 (1992) 249–261.
- Costerton JW, Stewart PS, Greenberg EP: Bacterial biofilms: a common cause of persistent infections. Science 284 (1999) 1318–1322.
- Wimpenny J: An overview of biofilms as functional communities. In: Allison D, Gilbert P, Lappin-Scott H, Wilson M (eds): Community structure and co-operation in biofilms, 1–24, Cambridge University Press, Cambridge (2000).
- Albertano P, Kovácik L, Marvan P, Grilli-Caiola M: A terrestrial epilithic diatom from Roman Catacombs. In: Marino D, Montresor M (eds), Proceedings of the Thirteenth International Diatom Symposium, 11–21, Bristol: Biopress Limited (1995).
- Hernández-Mariné M., Roldán M, Clavero E, Canals A, Ariño X: Phototrophic biofilm morphology in dim light. The case of the Puigmoltó sinkhole. Nova Hedwigia, Beih. 123 (2001) 237–253.
- Pantazidou A.: Cyanophytes (Cyanobacteria) found on the 5th century BC sculptures and inscriptions on the limestone walls of cave Nympholypton, Greece. In: Koutsoukos PG., Kontoyannis CG (eds), Proceedings of the 7th Eurocare-Euromarble Workshop. Patras: ICE/HT-FORTH, 153–57 (1997).
- Lawrence JR, Korber DR, Hoyle BD, Costerton JW and Caldwell DE: Optical sectioning of microbial biofilms. J Bacteriol 173 (1991) 6558–6567.
- Albertano P, Kovácik L, Grilli-Caiola M: Preliminary investigations on epilithic cyanophytes from a Roman Necropolis. Arch Hydrobiol/Suppl Algological studies 75 (1994) 71–74.