Pseudomonas, Pantoea and Cupriavidus isolates induce calcium carbonate precipitation for biorestoration of ornamental stone

Authors


Correspondence

Fotis Rigas, School of Chemical Engineering, National Technical University of Athens, 15700 Athens, Greece. E-mail: rigasf@central.ntua.gr

Abstract

Aims

Bacterially induced calcium carbonate precipitation from various isolates was investigated aiming at developing an environmentally friendly technique for ornamental stone protection and restoration.

Methods and Results

Micro-organisms isolated from stone samples and identified using 16S rDNA and biochemical tests promoted calcium carbonate precipitation in solid and novel liquid growth media. Biomineral morphology was studied on marble samples with scanning electron microscopy. Most isolates demonstrated specimen weight increase, covering partially or even completely the marble surfaces mainly with vaterite. The conditions under which vaterite precipitated and its stability throughout the experimental runs are presented.

Conclusions

A growth medium that facilitated bacterial growth of different species and promoted biomineralization was formulated. Most isolates induced biomineralization of CaCO3. Micro-organisms may actually be a milestone in the investigation of vaterite formation facilitating our understanding of geomicrobiological interactions. Pseudomonas, Pantoea and Cupriavidus strains could be candidates for bioconsolidation of ornamental stone protection.

Significance and Impact of the Study

Characterization of biomineralization capacity of different bacterial species improves understanding of the bacterially induced mineralization processes and enriches the list of candidates for biorestoration applications. Knowledge of biomineral morphology assists in differentiating mineral from biologically induced precipitates.

Introduction

Microbes have been interacting with metals and minerals since the Precambrian era (Ehrlich 1998). Biological processes involving calcium carbonate precipitation–dissolution preserve the Earth's neutral conditions in which most organisms thrive (Zavarzin 2002), and bacteria play an active role in both (Douglas 2005). Consequently, these processes can be classified under the term biomineralization and categorized as biologically controlled mineralization, such as bone development (Mann 2001) and biologically induced mineralization (BIM), where micro-organisms modify their local environment thus promoting the conditions for extracellular calcium carbonate precipitation (Gadd 2010). In BIM, bacterial cell surfaces and outer structures define the final outcome of the mineralization process (De Muynck et al. 2010a).

Exposed stone is subjected to physical, chemical and biological factors. Their combinations cause the dissolution of the mineral matrix rendering the material more porous and with decreased mechanical features (Tiano et al. 1999). Bacteria may induce dissolution of stone materials by excreting acidic compounds during their metabolic cycle. Conversely, they are capable of promoting calcium carbonate precipitation by influencing the factors related to precipitation, namely: Ca2+ concentration, dissolved inorganic carbon concentration, pH and presence of nucleation sites (Hammes and Verstraete 2002). Induction of calcium carbonate biomineralization proved to be a general bacterial phenomenon (Boquet et al. 1973) which has been consistently demonstrated. Thus, bacteria are considered as promising candidates for the development of environmentally friendly techniques aiming at the restoration of decaying stone monuments (Adolphe et al. 1990). For heterotrophic bacteria under aerobic conditions, the main metabolic pathways of interest are the oxidative deamination of amino acids, the degradation of urea and the utilization of organic acids (De Muynck et al. 2010a). In oxidative deamination, breakdown of amino acids releases NH4+ that raises pH levels and carbonate concentration, which can provoke calcium carbonate precipitation in the presence of calcium (Jimenez-Lopez et al. 2008). Alternatively, bacteria that possess the urease enzyme catalyse urea hydrolysis to carbonate and ammonium. Finally, organic acid assimilation assists in the bioprecipitation of calcium carbonate by increasing the concentration of carbonate during metabolism (Braissant et al. 2002). Increased calcium carbonate precipitation has been observed when the above-mentioned amino acid and organic acid pathways are simultaneously exploited (De Muynck et al. 2008a).

Bacillus species have been utilized due to their ureolytic abilities (Hammes et al. 2003). Castanier et al. (1999) utilized the heterotroph Bacillus cereus for the development of a biomineralization method that was applied successfully on walls and statues by spraying a suspension of the micro-organism first and subsequently feeding it with the appropriate medium. Bacillus sphaericus was initially applied on limestone (Dick et al. 2006) and subsequently on concrete (De Muynck et al. 2008a; Van Tittelboom et al. 2010). For the latter substrate, Sporosarcina pasteurii (formerly Bacillus pasteurii) was tested on concrete cracks (Bang et al. 2001; Sung-Jin et al. 2010) and recently on bricks (Sarda et al. 2009). Rodriguez-Navarro et al. (2003) suggested the use of Myxococcus xanthus as another appropriate bacterium, which indeed promoted calcite and vaterite formation.

The goal of this work was to investigate bacterial species obtained from the surfaces of marble taken from the ancient quarry site of Pentelic marble. Nonfastidious micro-organisms were isolated, and the growth medium was optimized for improving their efficiency for calcium carbonate precipitation. Experiments with the selected culture medium dosed with different concentrations were performed to identify the calcium carbonate polymorphs that prevailed and to distinguish the isolates that could be considered as candidates for bioconsolidation strategies.

Materials and methods

Isolation of micro-organisms

Micro-organisms were isolated from specimens taken from an ancient marble quarry, in the Penteli Mountain, Athens, Greece. Initially, the surface of the marble was wiped with a cotton swab, previously immersed in nutrient broth medium (Merck KGaA, Darmstadt, Germany) and then resuspended in the medium. Additionally, using a sterile blade, stone samples were scraped into a sterile test tube. All samples were cold-stored in an icebox during transport to the laboratory where they were immediately processed.

Stone samples weighing 0·1 g were added in a 0·9% (w/v) NaCl −0·1% (w/v) Triton-X solution, vortexed for 5 min, and 50 μl aliquots were inoculated in the nutrient broth (Merck). All media were supplemented with nystatin (kindly provided from Bristol Myers Squibb, Athens, Greece) to inhibit the growth of fungi and then incubated at 28°C/120 rpm on a horizontal shaker (Labnet Orbit-1000; Edison, NJ, USA) for 72 h. Nutrient broth medium inoculated on-site with a cotton swab was supplemented with nystatin and similarly incubated.

Every 24 h, an aliquot was serially diluted and spread on plates of the corresponding medium. Individual colonies were subcultured in liquid media and serially diluted until uniform colony morphology was observed. After purification, bacteria were stored at −30°C in PROTECT vials – bacterial preservers (Technical Service Consultants Limited, Lancashire, UK) according to the manufacturer's instructions.

Molecular biology identification

Isolation of bacterial DNA was performed by diluting bacterial cells, recovered from colonies grown in nutrient broth agar plates for 18 h, by gently touching the colony surface with a sterile pipette tip and immersing the tip in 50 μl of ultra pure water (Merck). Subsequently, samples were heat shocked at 98°C and centrifuged at 2225 g for 2 min. Next, 2 μl of the supernatant was used for 16S rDNA amplification with the eubacterial universal primers: 27 forward 5′-AGAGTTTGATCMTGGCTCAG-3′ and 1492 reverse 5′-TACGGYTACCTTGTTACGACTT-3′ in the following reaction concentrations: 1× PCR buffer, 2 mmol l−1 MgCl2, 0·166 mmol l−1 dNTPs, 0·233 μmol l−1 of each forward and reverse primer and 1·5 U Taq polymerase (Gennaxon) to a final volume of 30 μl. The thermal profile of the PCR reactions was as follows: initial denaturation at 94°C for 3 min, followed by 35 cycles of 94°C for 1 min, 48°C for 1 min and 72°C for 3 min and a final cycle at 72°C for 10 min. The length of the amplified 16S rDNA fragment was approximately 1500 bp. The amplified fragments were then purified with the QIAquick PCR purification kit (QIAGEN, Hilden, Germany), cloned with a TOPO TA Cloning Kit and transformed into chemically competent cells in accordance with the manufacturer's instructions (Life Technologies,Carlsbad, CA, USA). Transformants were selected on LB amp + Xgal and screened by PCR using vector primers T7 and SP6 under the following conditions: 3 μl of the lysed bacteria, 1× PCR buffer, 1·5 mmol l−1 MgCl2, 0·15 mmol l−1 dNTPs, 1·5 U Taq polymerase and in a final volume of 20 μl. Reaction mixtures were incubated at 94°C for 5 min as an initial incubation step followed by 30 cycles at 94°C for 1 min, 50°C for 1 min, 72°C for 3 min and then a final step at the same temperature for 10 min.

The resulting PCR products were purified using QIAquick PCR purification kit, and sequences were determined by the DNA analysis facility at Yale University using T7 and SP6 primers.

Nearly full 16S rDNA was sequenced for the isolated micro-organisms and compared with the NCBI database with BLASTN network service (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The sequences were submitted to GenBank, and their corresponding accession numbers are presented in Table 1.

Table 1. Phylogenetic affinities of 16S rRNA gene sequences of stone isolates
Isolate nameSample origin16S rDNA length (bp)NCBI closest matchPercentage of similarity (%)Accession NumberFamilyClassPhylumaRisk group
  1. a

    Risk Group: The category that each isolate corresponds to is based on the type of strains deposited in the German Collection of Micro-organisms and Cell Cultures (DSMZ) and the American Type Culture Collection (ATCC), and references therein. Risk Group 1 micro-organisms are not known to cause disease in healthy adult humans. Risk Group 2 micro-organisms present a moderate risk and should be handled under specific guidelines.

PRStone/cotton swab1504Pantoea agglomerans/vagans99·80/99·53 JQ412064 EnterobacteriaceaeGammaproteobacteriaProteobacteria2
B_MPEZStone1496 Cupriavidus metallidurans 99·87 JQ046372 BurkholderiaceaeBetaproteobacteriaProteobacteria1
BLStone1520 Paenibacillus polymyxa 98·95 JQ412065 PaenibacillaceaeBacilliFirmicutes1
CStone1515 Bacillus cereus 99·54 JQ412066 BacillaceaeBacilliFirmicutes2 (DSMZ)/1(ATCC)
T5_LStone1501 Pseudomonas chlororaphis 99·53 JQ412067 PseudomonadaceaeGammaproteobacteriaProteobacteria1
T10Stone1513 Bacillus licheniformis 99·80 JQ412068 BacillaceaeBacilliFirmicutes1

Sequences were aligned using default parameters in CLUSTAL W and corrected by eye. Sequence divergences were estimated with the software MEGA v. 5 (www.megasoftware.net) (Tamura et al. 2011) choosing Kimuras' two-parameter model (Kimura 1980) of evolution, and phylogenetic trees were constructed using the neighbour-joining algorithm (Saitou and Nei 1987). Branch support was assessed with 1000 bootstrap replicates. Sequences for construction of the trees, according to the results of the BLASTN program, were selected based on:

  1. The closest type of strain that presented 100 or 99% similarity,
  2. The length of the 16S rDNA compared should have been equal to or close to the length of our isolates,
  3. The maximum score which should be as high as possible,
  4. Sequences corresponded to cultured micro-organisms

For example, the phylogenetic tree for B_MPEZ isolate was constructed based upon the above prerequisites and from similar isolates that facilitated the strain's initial identification and latter affiliation to the genus Cupriavidus (Goris et al. 2001; Vandamme and Coenye 2004). Pandoraea apista strain LMG 20576, Burkholderia cepacia strain ATCC25416 and Alcaligenes faecalis strain TZQ4 served as out-groups (see Fig. S1).

Phenotypic tests

API 20NE, 20E and CHB/E test strips (bioMérieux, Marcy l'Etoile, France) were used to visualize biochemical characteristics and assist in identification. Gram stain, oxidase and catalase tests were separately evaluated. Results were compared with the API Identification Tables and corresponding literature.

Growth media

B4 medium (Boquet et al. 1973) for biomineralization experiments was modified. Glucose was omitted, and instead of yeast extract different nutrient sources were investigated; tryptose (LABM, Lancashire, UK and Fluka; Sigma-Aldrich, Munich, Germany), bacteriological peptone (BP), soy peptone, beef extract and proteose peptone A (LABM, Lancashire, UK). Ca(CH3COO)2 (Sigma-Aldrich, Munich, Germany) solution remained the source of calcium and potential enhancer of bacterial growth through the assimilation of acetate. Before sterilization, the pH was adjusted to 8 with NaOH. Qualitative observation of growth was carried out by monitoring absorbance at 620 nm with a Hach DR-2000 spectrophotometer (HACH LANGE GmbH, Düsseldorf, Germany) in samples diluted with water (1 : 1 dilution).

Stone substrate

Marble from the Dionysus site in the Penteli Mountain was the selected stone in the biomineralization experiments. This is the same type of marble that the renowned ancient monuments in Athens, like the Parthenon, were constructed with. Stone samples were cut in sizes of 4 cm × 1 cm × 1–1·5 cm. Initially, they were cleaned of soil debris with tap water, then rinsed thoroughly with deionized water and allowed to dry at 50°C for 24 h. Stone slabs were sterilized in the same manner as growth media and weighed prior to immersion in the growth medium.

Biomineralization experiments

The growth medium that proved most appropriate was 0·36 g BP and 2·5 ml of 10% (w/v) Ca(CH3COO)2 in 100 ml of deionized water (BP1x). BP agar media were prepared at 1×, 2× (0·72 g BP and 5 ml of 10% (w/v) calcium acetate) and 4× (1·44 g BP and 10 ml of 10% (w/v) calcium acetate) concentrations. Sterile stone samples were immersed in 99 ml of the above liquid medium for the biomineralization experiments and inoculated with 1 ml of each micro-organism, pregrown for 24 h. Control samples (without micro-organism) were used in all experiments and subjected to the same treatment. Additional series of control samples with dead bacteria were incubated for 30 days to investigate whether bacterial constituents could provide a template for calcium carbonate precipitation. Flasks were incubated at 30°C on a horizontal shaker at a speed of 120 rpm for 15 or 20 days. Every 48 h, a loopfull from each flask was applied on corresponding agar media to test for contamination.

For sample processing, 6 ml of medium was centrifuged to obtain the supernatant for further analysis. Viability and enumeration of bacteria were performed with the serial dilution method in 0·1% (w/v) peptone from casein (Merck) solution, absorbance of the medium was monitored at 620 nm, and the pH was monitored with a WTW Inolab pH-720 pH meter (WTW GmbH, Weilheim, Germany). The rest of the medium was filtered, and biomineral attached to flask walls was gently scraped off to obtain adequate amounts of calcium carbonate for further analyses.

Recovered stone samples were rinsed thoroughly with deionized water and dried at 50°C. After 48 h, they were weighed, and the final weight gain was calculated against the weight of the sample prior to its immersion in the growth medium. Subsequently, samples were prepared for scanning electron microscopy analysis.

Calcium and acetate monitoring

Calcium depletion was monitored with atomic absorption spectroscopy using a Varian spectrophotometer (model AA240FS). The wavelength for calcium detection was at 422·7 nm. Acetate assimilation was monitored with liquid chromatography in a DIONEX BioLC (Dionex Thermo Fisher GmbH, Idstein, Germany) coupled with an ED50 electrochemical detector and a GS50 gradient pump. The column used was an IonPac AS14 analytical. Data were recorded with Chromeleon 6.20 sp2 Build 541 software (Dionex Thermo Fisher GmbH, Idstein, Germany).

Scanning electron microscopy and ultrasonic treatment

Stone samples were first coated with platinum by electrodeposition and then placed under a JEOL scanning electron microscope to observe the morphology of the induced biomineral. Ultrasonic treatment of biomineralized samples was carried out, as suggested by Rodriguez-Navarro et al. (2003). Briefly, biomineralized and blank stone samples were immersed in deionized water in an upright position (one of their small sides 1 cm × 1–1·5 cm was touching the bottom of the glass beaker). Sonication occurred for 5 min, and samples were immediately rinsed with deionized water and dried at 50°C for 24 h prior to weighing. The process was performed five times.

XRD and FT-IR analyses

Crystals isolated from the biomineralization media were collected, finely powdered, resuspended in Milli-Q water, filtered through silver membrane filters (pore size 0·45 μm) and subjected to X-ray diffraction. The instrument used was a Rigaku D/MAX B diffractometer (Rigaku Europe SE, Ettlingen, Germany), with CuKa radiation, and graphite monochromator, operated at 40 kV, 20 mA. The scan range was 2·0–50·0°2θ, in steps of 0·01°.

Fourier transform infrared spectroscopy (FT-IR) analysis was performed in all samples isolated to have a prompt result regarding the biomineral. A few crystals were mixed with KBr (Merck for spectroscopy) and pulverized in an agate mortar to form a homogenous powder from which, under a pressure of 7 tons, the appropriate pellet was prepared. All spectra were recorded from 4000 to 400 cm−1 using the Bio-Rad FTS 3000 MX spectrometer (Bio-Rad Laboratories Ltd, Hemel Hempstead, UK). Scans were 32 per spectrum with a resolution of 4 cm−1. The IR spectra were analysed using the spectroscopic software Win-IR Pro Version 3.0 (Bio-Rad Laboratories Ltd, Hemel Hempstead, UK) with a peak sensitivity of 2 cm−1.

Results

Isolation, purification and identification of isolated micro-organisms

The isolation strategy followed in this work aimed to culture nonfastidious micro-organisms inhabiting stone surface. Isolates with their corresponding phylogenetic identification are presented in Table 1. All bacteria were identified from 16S rDNA sequencing, and their phylogenetic trees proved to be consistent in results. Isolates, with the exception of Paenibacillus polymyxa, were able to grow in BP, tryptose and yeast extract agar plates. Temperature growth range was 20–37°C with no growth above 40°C or below 10°C. API results assisted in bacterial identification only in accordance with DNA sequencing.

Bacillus cereus C, Bacillus licheniformis T10, Ppolymyxa BL and C  metallidurans B_MPEZ provided results similar to the corresponding API Identification Tables, Bergey's manual of systematic bacteriology (Logan and De Vos 2009) and corresponding literature (Goris et al. 2001; Vandamme and Coenye 2004; Vaneechoutte et al. 2004). For T5_L isolate, the phenotypic tests coincided with Pseudomonas chlororaphis but could not differentiate our isolate in the subspecies level (Peix et al. 2007) (Table S1).

Finally, for the isolate designated as PR, according to API tests, it was identified as Pantoea sp. Genetic identification provided us with two results corresponding with Pantoea agglomerans and Pantoea vagans. Further differentiation of the species level could not be performed because of differences in the biochemical tests (Table S2). If Pvagans (Brady et al. 2009) was not recently isolated, our micro-organism would be affiliated adequately to Pagglomerans (Delétoile et al. 2009). Owing to this ambiguity, it is designated as Pantoea sp. PR.

Growth media

B4 biomineralization medium was differentiated due to the fact that when components were mixed prior to, or after being sterilized, a cloudy solution formed a few hours after pH adjustment (pH = 8) leading to a fine precipitate that could not be isolated by filtration.

Yeast extract replacement was based on calculations for nitrogen content that were provided by the manufacturer. The goal was a medium that would be able to induce similarly good growth rates with an adjustable pH (a pH value close to 8) without inorganic precipitation. The best results were obtained with BP. Most micro-organisms grew in this medium with no delay, even from frozen stock cultures (as identified by measuring the absorbance of inoculated growth media after 18 h). BP's pH was adjusted prior to sterilization and was mixed with calcium acetate after being separately sterilized.

All micro-organisms except Ppolymyxa BL were able to grow in BP medium even at the lowest concentrations (BP1x). When concentration of BP media was doubled (BP2x), bacteria underwent a period of latency, followed by a period when nutrients from BP were metabolized. Acetate further promoted bacterial growth by the diauxic effect for all isolates except Pantoea sp. PR, whose absorbance values during growth suggested that it was readily assimilated since the start of its growth. Acetate assimilation coincided with the onset of detectable calcium carbonate formation.

Biomineralization on marble surface

Initial screening of biomineralization was performed at various concentrations of BP solid media. As shown in Table 2, all micro-organisms at all concentrations except at BP4x formed calcite in solid media as the main polymorph (analysis by FTIR). As presented in Fig. 1, polymorphism could be monitored by FT-IR and XRD.

Table 2. Calcium carbonate polymorphism of micro-organisms under different solid and liquid media concentrations
IsolateSolid BP1xSolid BP2xSolid BP4xLiquid BP1xLiquid BP2xLiquid BP4xCrystal morphologyCrystal colour
  1. Calc, calcite; Vat, vaterite.

  2. Polymorphs in parentheses correspond to minor but detectable amounts.

Pantoea sp. PRCalcCalc+(Vat)VatCalcVatVatMineralized bacteria; spheresLight yellow
Cupriavidus metallidurans B_MPEZCalcCalcCalcCalcVat+(Calc)VatMicritic spheres; short rods; combination of spheres and rods.White-beige
Bacillus cereus CCalcCalcVatCalcVat+(Calc)VatMineralized bacteria (vaterite); dendritic (calcite).Brown
Pseudomonas chlororaphis T5_LCalcCalc+(Vat)VatCalcVatVatMicritic spheresWhite-green crystals
Bacillus licheniformis T10CalcCalc+(Vat)VatCalcVatVatMineralized bacteriaBrown
Figure 1.

FTIR and XRD spectra of calcite (a and b) and vaterite (b and c) identified in biomineralization experiments. Calcite peaks on FTIR are at 713, 875, 1423 (single), 1794 and 2516 cm−1. Vaterite peaks are at 745, 875, 1084, 1437–1487 (double peak) and 2508 cm−1. Polymorphism was further investigated with XRD, where the corresponding 2θ degrees for calcite were at 23·05 (9%), 29·40 (100%) and 35·96 (15%) and for vaterite were at 24·87 (90%), 27·01 (100%) and 32·80 (90%).

Biomineral polymorphism differed in liquid media. Calcite was detected at BP1x, and vaterite was observed at concentrations BP2x and higher (Table 2). At concentration BP1x, biomineralization was very low, especially on marble, and the actual morphology of the biomineral polymorphs could not be easily distinguished from marble fragments. All tested micro-organisms provided better surface biomineralization at BP2x than BP4x. Furthermore, weight increase was not analogous to the nutrient surplus (increase from BP2x to BP4x) provided, and micro-organisms tended to coagulate and form large biomineral assemblies instead of a smoother calcium carbonate surface on the substrate. This phenomenon was not associated with nutrient assimilation; at the end of the experiments, acetate and calcium concentrations were either very low or insignificant under all investigated growth medium concentrations.

Bioconsolidation strategies for preservation of stone monuments and works of art require knowledge of the biomineral's morphology in liquid media for each isolate. Such information could assist in distinguishing bioprecipitation from inorganic formation of calcium carbonate in an in situ application. Furthermore, sufficient morphological data are useful in determining the potential of a micro-organism for use in such preservation strategies. Marble samples from BP2x and BP4x liquid cultures analysed with scanning electron microscopy revealed similar morphologies at both concentrations for each isolate (Table 2).

To elucidate the biomineral's morphology and its polymorph correspondence, samples of the newly formed calcium carbonate filtered from the medium for each isolate were analysed with FT-IR and XRD. Furthermore, these samples were also examined under the electron microscope. Electron microscopy photographs (data not shown) revealed a biomineralization morphology similar to the one observed in marble samples. No precipitation was detected on blank experiments regardless of the duration of incubation and medium concentration. Morphology of blank marble samples is shown in Fig. 2a–c. The different orientations of the calcium carbonate filaments that comprise the substrate are presented under low magnification (Fig. 2a). Increased resolution elucidates marble fragments due to the substrate's cutting process (Fig. 2b–c). Fragments show a polyhedral morphology (Fig. 2c) discerning them from biomineralization. Sterilization of the marble did not induce any considerable fragmentation of the surface.

Figure 2.

Biomineral morphologies induced by isolated bacteria. (a–c) blank marble; (d, e) Pantoea sp. PR: a general view under low magnification of the marble surface is presented in (d) where both large spherical bacterial assemblies and a layer of biomineralized bacteria coexist. Magnification of the biomineralized bacteria is shown in (e); (f) P. chlororaphis T5_L novel biomineral surface constituted of micritic spheres; (g, h) Blicheniformis T10: general view of the marble biomineralized surface in (g) and biomineralized bacterial rods consisting of the novel surface in (h); (i, j) Bcereus C showing different morphologies on different areas of the same sample that correspond to vaterite as biomineralized cells of rod shape (i). Gradual expansion of the novel biomineral covers the marble surface. Uncovered area is at the right edge of the picture. Calcite (j) was observed creating dentritic structures; and (k, l) Cmetallidurans B_MPEZ: overview of the marble sample covered completely from the novel biomineral (k), whose morphology is spheres of different diameter (l).

Pantoea sp. PR biomineral consisted of scattered large spherical assemblies and a smoother novel surface covering the marble (Fig. 2d). Mineralized bacteria created the large spherical morphologies by nucleating in larger numbers compared with the rest of the marble surface. The sphere's surface and the novel layer covering the rest of the substrate were consisted of biomineralized bacterial cells as shown in Fig. 2e. FT-IR analysis showed that samples corresponded to vaterite.

Pseudomonas chlororaphis T5_L covered the surface of the marble sample with micritic spherical morphologies (Fig. 2f). Biomineralized thin rods could be observed scattered between the spheres suggesting that these morphologies were developed in close relation to microbial utilization of calcium and induction of bioprecipitation.

Bacillus licheniformis T10 covered the substrate with a novel vaterite layer creating an uneven biomineral surface (Fig. 2g), consisted of mineralized bacteria of a size corresponding with the micro-organism's dimensions (Fig. 2h). On the contrary, Bcereus C showed a mixture of vaterite and calcite morphologies in different areas of the same sample. In Fig. 3i,j, the different morphologies are presented. Vaterite biomineral showed itself as biomineralized bacteria (Fig. 2i). On the contrary, crystals assembled together creating dendritic morphology (Castanier et al. 1999; Ben Chekroun et al. 2004) corresponded to calcite (Fig. 2j).

Figure 3.

(a) Graphical presentation of weight increase from each isolate during the 20 or 15 or 20-day experiments in BP2x media. Micro-organisms: (■) Blicheniformis T10, (▲) Pantoea sp. PR, (×) Bcereus C, (●) C. metallidurans B_MPEZ, (♦) P. chlororaphis T5_L and (*) blank marble; (b) Weight loss of biomineralized marble samples of the corresponding micro-organisms during sonication treatment. Samples selected for sonication treatment correspond to those with the highest weight increase through the experimental period, namely: Cmetallidurans B_MPEZ from day-10, Pantoea sp. PR from day-15, Blicheniformis T10 from day-15 and Pchlororaphis T5_L from day-15. Bcereus C was not tested as this strain has already been used by another research group (Castanier et al. 1995). Micro-organisms: (♦) P. chlororaphis T5_L, (■)Blicheniformis T10, (▲) Pantoea sp. PR, (●) C. metallidurans B_MPEZ and (×) blank marble. Standard deviation bars are shown.

Finally, Cmetallidurans B_MPEZ succeeded in covering the marble surface (Fig. 2k) with morphologies consisted of micritic spheres, short rods and combinations of spheres and rods (Fig. 2l).

Further experiments were conducted in BP2x media for each micro-organism with samples obtained every 5 days for 15- or 20-day tests depending on their viability. Controls without micro-organisms were simultaneously analysed; no inorganic precipitation of calcium carbonate was observed nor was there any contamination due to the long incubation period. Figure 3a shows the weight increase during these tests. Negative values of blank samples correspond to minor dissolution of the marble surface in the experimental solution. Pantoea sp. PR provided promising results regarding biomineralization on marble surfaces with a gradual increase in sample weight and coverage. Full coverage of the marble substrate with a layer of light yellow-coloured vaterite was detected after 10 days. Loss of the micro-organisms' viability was observed after 15 days, affecting the morphology of the final sample, where filaments of the biomineral adhered to the marble's surface. Bacillus cereus C showed the highest weight increase. The solution of the medium was brown, affecting the calcium carbonate's colour and presenting a considerable aesthetic difference to the naked eye. Bacillus licheniformis T10 did not achieve considerable biomineralization. Despite its detectable weight increase, vaterite was brown in colour due to the bacterial pigments in the solution observed after the third day. Both Bacillus strains remained viable until the end of the experiments (20 days).

Pseudomonas chlororaphis T5_L covered the entire surface of the marble with a smooth coating of white-green crystals after 10 days. The colour gradually changed to grey-green due to consecutive biomineral layers. The micro-organism remained viable until the end of the test.

Finally, C. metallidurans B_MPEZ provided gradual substrate coverage with a uniform white-beige vaterite showing an optimum weight increase after 10 days. The colour of the solution was white throughout the experiment, and the micro-organism remained viable.

The pH of the final solution in the experiments was 9·0–9·3 which assisted in the bioprecipitation of calcium carbonate. Acetate was exhausted from all micro-organisms by the tenth day of the experiments. A decrease in dissolved calcium was observed only after initiation of acetate assimilation, which kept diminishing until the fifteenth day of the experiments.

Samples were subjected to sonication treatment. Sonication forces act upon all sides of the marble sample, while the surface with the new calcium carbonate layer was the one facing the medium (4 × 1 cm) during biomineralization experiments. Therefore, the overall weight loss does not correspond solely to the latter surface. Biomineralized marble substrates showed a detectable decrease in weight loss compared with blanks proving that the new calcium carbonate layer on the substrate was coherent enough to withstand sonication forces and protect the marble underneath (Fig. 3b). Pseudomonas chlororaphis T5_L presented the highest resistance to sonication forces compared with blank marble. Pantoea sp. PR came second, while Cmetallidurans B_MPEZ weight loss was similar to the decrease in blank marble. Bacillus licheniformis T10 (used for comparison purposes due to its reported biomineralization abilities) was in the middle of the weight loss under 5 consecutive 5 min sonication tests. In all cases, the overall weight loss was lower than the amount of biomineral deposited on the substrate's surface.

Discussion

Evaluation of a biomineral layer on stone substrate by different isolates was performed. Pentelic marble was selected as the test template because it has been the construction material of choice for some of the most renowned ancient monuments in Greece (e.g. the Parthenon). Monuments and statues from ancient Athens and other locations such as Delphi and Delos were constructed with Pentelic marble or marbles with similar physicochemical characteristics (Parian marble from Paros Island). One of their main characteristics is their low to negligible porosity. It is possible that this may be one of the factors why those monuments still stand in an open air environment. Low porosity provides a hostile environment for bacteria due to the fact that there are no sites for easy colonization.

Appropriate identification of all isolates was achieved by the isolation and sequencing of almost the full length of 16S rDNA as was advised previously (Rappé and Giovannoni 2003). API tests that were used to verify the phenotypic characteristics of the micro-organisms proved similar according to the literature for each micro-organism. Nevertheless, identification of the species level varied depending on the isolate. Bacillus, Paenibacillus and Cupriavidus could be identified at the species level, but Pantoea (species level) and Pseudomonas (subspecies level) isolates with very close biochemical and phenotypical characteristics could not be differentiated. It is preferable to initially perform genetic identification, followed by accurate biochemical tests.

Pantoea sp. PR could not be unambiguously differentiated in the species level among Pagglomerans and a recent novel isolate Pvagans (Brady et al. 2009). Regardless of this, P. agglomerans (Gavini et al. 1989) is a micro-organism isolated from plants, soil, water and humans. It is considered a plant pathogen (Brady et al. 2008), but its antagonistic activity against other plant pathogens resulted in its application as a biological control agent (Nunes et al. 2002; Jin et al. 2003; Völksch et al. 2009). In addition, it has been linked to human nosocomial infections (Brady et al. 2008; Völksch et al. 2009). Nevertheless, its biotechnological applicability does not restrict its evaluation against other applications, as long as precautions are taken (Table 1). Biochemical differences of the isolate in this study and those reported in the literature outlined that a more comprehensive survey should be carried out in case such a micro-organism is selected for further investigation. It is the first time, to the best of our knowledge, that such an isolate has been evaluated for its ability to induce calcium carbonate precipitation.

Cupriavidus metallidurans B_MPEZ (formerly Ralstonia) was isolated from the surface of exposed marble in the Penteli Mountain, a site that is not so far regarded as heavily polluted by heavy metals. This confirmed the fact that C. metallidurans is a micro-organism that may be found in different environments, although initially isolated from industrial sites in Belgium during an excavation for metal-resistant bacteria (Goris et al. 2001). Its ability to withstand and survive in biotopes suffering from high heavy metal concentrations resulted in its full genome sequence (Janssen et al. 2010) and its ongoing research regarding the micro-organism's exploitation in bioremediation applications. Its ability to proliferate under toxic conditions of heavy metals could prove to become an advantage for monuments in urban environment. Cupriavidus metallidurans B_MPEZ provided coverage of the substrate with an acceptable aesthetic effect for the naked eye and presented a uniform surface.

Bacillus cereus has been applied by Le Métayer-Levrel et al. (1999), who were among the first to develop a methodology named Calcite Bioconcept technique (Adolphe et al. 1990) for monument protection. After its identification, the isolate was utilized in our experiments during screening to verify the ability of the proposed growth medium to assist in biologically induced calcium carbonate precipitation. In general, Bcereus and Blicheniformis have been detected in numerous monument sites worldwide (Zanardini et al. 1997; Laiz et al. 1999; Videla et al. 2000; Gorbushina et al. 2004; McNamara et al. 2006; Nuhoglu et al. 2006; Scheerer et al. 2009; Pepe et al. 2010) and cave environments (Baskar et al. 2009; Banks et al. 2010).

Pseudomonas species have been detected as part of the microflora in caves with Bacillus species (Laiz et al. 1999; Groth et al. 2000), on stone monuments (Urzi et al. 1999; Videla et al. 2000; Gorbushina et al. 2002; McNamara et al. 2006; Jimenez-Lopez et al. 2007; Scheerer et al. 2009; Jroundi et al. 2010) and wall paintings (Gurtner et al. 2000). Their identification was also reported on palaeolithic paintings with Ralstonia (Schabereiter-Gurtner et al. 2002). Biotechnological approaches have already used Pseudomonas endogenous capabilities for biorestoration of frescoes (Ranalli et al. 2005), and Zamarreño et al. (2009) presented the potentials of Pseudomonas putida and Pseudomonas aeruginosa. Our isolate promoted biomineralization of calcium carbonate and was the isolate that resisted most in sonication forces having the lowest weight loss (Fig. 3b).

To develop an alternative medium, a nonfastidious nutrient solution with the least components possible was considered. An additional issue that should be considered further is that the colour of the solution varied from light to dark yellow depending on medium concentration. This had an adverse effect on some isolates with colour transferring to the biomineral, especially in higher medium concentrations. Inducing colour differentiations is not desirable, unless the micro-organism inevitably produces pigments which may be exploited in bioconservation interventions similarly to Le Métayer-Levrel et al. (1999).

Most of the isolates induced calcium carbonate precipitation establishing themselves as potential candidates for bioconsolidation methodologies (Zamarreño et al. 2009). The time frame in which bioprecipitation was detected differed among isolates presenting their different metabolic response through the diauxic effect of acetate. Bcereus C and Pchlororaphis T5_L required a longer period to trigger acetate assimilation. Thus, calcium carbonate precipitation was detected after the fifth day of incubation. On the contrary, Pantoea sp. PR, Cmetallidurans B_MPEZ and to a lesser extent Blicheniformis T10 were able to utilize acetate earlier showing a weight increase from the fifth day (Fig. 3a). Despite the delay, Pchlororaphis T5_L biomineral proved to confront sonication forces with the minimum overall weight loss compared with the rest of the isolates. Cmetallidurans B_MPEZ provided the least protection.

In this work, heterotrophs and their oxidative deamination of amino acids metabolic pathway were exploited. This pathway, along with the degradation of urea, is generally utilized to achieve the prerequisites for microbial carbonate precipitation: pH increase and CO32− production. For a financially feasible method, selective media are not favourable, although they could be more efficient in the case of a specific bacterium. Therefore, the use of yeast extract and peptones is promoted to provide the appropriate elements for enhancement of bacterially induced mineralization (Boquet et al. 1973; Adolphe et al. 1990; Castanier et al. 1995; Rodriguez-Navarro et al. 2003). From the nitrogen sources investigated, BP provided the most satisfactory results. Despite being less nutritious than yeast extract, it facilitated a pH increase close to 8 without the formation of a cloudy solution. Calcium remained soluble and did not precipitate requiring filtration. Microbial metabolism facilitated pH increase to promote biomineralization. The pH increase was assisted by the oxidative deamination of BPs components and ammonium release (Gottschalk 1985; Castanier et al. 1999; Hammes and Verstraete 2002).

display math(1)

Acetate could hinder this pH increase by acting as a buffer. Its assimilation eliminated this characteristic allowing solution pH to reach values of 9·0–9·3. The creation of an alkaline environment has an impact when carbonate becomes part of the equation which is produced simultaneously during degradation of proteins and peptides (Gottschalk 1985).

In the medium applied in this study, glucose was omitted to avoid the production of acids that could dissolve the substrate or even the newly formed calcium carbonate. It has been reported that glucose as a supplement of a peptone solution inhibited biomineralization (Portillo et al. 2009). Acetate, an additional carbon source in the medium, is directly mineralized to carbon dioxide (assimilation of organic acids) by numerous aerobic bacteria, thus boosting microbially induced calcium carbonate precipitation (Portillo et al. 2009). This was observed in all our isolates. Acetate is assimilated from heterotrophs by integrating the compound in the tricarboxylic acid cycle (Braissant et al. 2002) assisting in their proliferation – cell density increase (diauxic effect) and higher CO2 concentration. Consequently, their microenvironment close to the cell membrane differs from the increased production of carbon dioxide which, in aqueous solution, transforms to carbonate ions.

display math(2)

Equation (1) presents the pathway for pH increase. This reduces the concentration of H+ produced in eqn (2) shifting the equilibrium towards carbonates. In such an alkaline pH, calcium carbonate formation is favoured. Under our experimental conditions, increased cell density speeded up the process via its metabolic products and cell membrane nucleation sites. This further affected the final polymorph that precipitated, which in this work, was mainly vaterite.

Vaterite is one of the three anhydrous polymorphs of calcium carbonate with aragonite and calcite comprising the rest. Vaterite transforms to the most stable calcite under ambient conditions of pressure and temperature (Suzuki et al. 2006). Samples analysed with FT-IR, XRD and SEM showed that polymorphism could correlate with the morphology of the biomineral. Calcite was observed in solid media and in low liquid medium concentrations. Stability of calcite under similar experimental conditions has been presented by other research groups in different media (Buczynski and Chafetz 1990; Tiano et al. 1999; El Kahoui et al. 2000; Hammes et al. 2003; Ben Chekroun et al. 2004; Cacchio et al. 2004; Dick et al. 2006; Baskar et al. 2009). The slow growth rate of micro-organisms (solid and liquid media) and diffusion of nutrients and metabolic products through the colony promoted the formation of calcite. This time frame accommodated the transformation of any intermediate biomineral to the most stable polymorph, or directly to its final formation. Calcite morphology, presented in Fig. 2j, is similar to the ones presented by Ben Chekroun et al. (2004). On the contrary, vaterite was consistently detected with the increase in nutrient concentration (Table 2). When bacteria were grown in solution, vaterite presented a consistent spherical morphology of various sizes, including structures that corresponded to bacterial dimensions (Table 2, Fig. 2). Analogous morphologies have been reported for different micro-organisms (Rodriguez-Navarro et al. 2003, 2007; Ben Chekroun et al. 2004; De Muynck et al. 2008a,b, 2010b). Spherical morphologies can be inferred to be vaterite, while rhombohedral morphology denotes calcite (Jimenez-Lopez et al. 2008). Constant detection of vaterite in higher concentrations in liquid (BP2× and 4×) media outlines that nutrient surplus and increased cell growth promoted vaterite's stability. High cell growth provided the template (cell walls) where calcium cations imminently interacted with negatively charged cell walls creating a saturated environment regarding calcium. Simultaneously, increased metabolism provided the appropriate pH increase, which further assists in the deprotonation of acidic moieties of the cell walls and carbonate production. The latter, once diffused out of the cell coordinated with calcium for the formation of the biomineral. Owing to nonequilibrium conditions (increased pH, calcium and carbonate concentration) close to the cells, the Ostwald's rule of stages (Mann 2001) is, in effect, kinetically favouring vaterite. Vaterite's transformation to the most stable calcite is inhibited by the incorporation of organic compounds that became entangled during precipitation (Xiao et al. 2010; Natoli et al. 2010) in close vicinity of the cells or from the cell wall itself. Our results showed that vaterite morphology was observed as spheres of various sizes always closely related to biomineralized bacteria. Finally, XRD and FT-IR analysis of samples after 1 year of isolation showed the same spectra as the ones presented in Fig. 1.

The nutrient medium developed in this work presents similarities to the ones previously reported by other groups (Rodriguez-Navarro et al. 2003; De Muynck et al. 2010b), that is, calcium and acetate concentration and nitrogen source. There could be an effect on mineralogy based on the nutrient's components. For example, we did not use a buffer solution, inducing the formation of supersaturated conditions and finally to vaterite precipitation, similarly to Rodriguez-Navarro et al. (2003). Acetate is an important factor of bacterial growth under our conditions and assists in the final morphology of the precipitate as reported by De Muynck et al. 2008a. Further investigation regarding our isolates to delineate the effect of the growth components is currently being investigated.

Cupriavidus metallidurans B_MPEZ, Pantoea sp. PR and Pseudomonas chlororaphis T5_L could be further investigated to optimize biomineralization conditions. The results presented here showed that each micro-organism has its own advantages and disadvantages regarding biomineralization efficiency, coherence to the substrate, colour impact on stone and biomineral morphology. A more detailed investigation regarding the time for biomineralization, nutrient assimilation and calcium carbonate bioprecipitation under conditions that resemble the ones in situ is currently under way.

Acknowledgements

This work was supported by the Ministry of National Education and Religious Affairs (Community Support Framework 2000–2006) under the Pythagoras II research programme. The project is co-funded by the European Social Fund (75%) and National Resources (25%) We would like to thank Dr. Aglaia Antoniou and Katerina Skaraki for their guidance and assistance in molecular biology experiments and Dr. Konstantina Papadopoulou for her assistance in sonication experiments.

Ancillary