PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Barton, Hazel]On: 10 June 2010Access details: Access Details: [subscription number 922886987]Publisher Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Geomicrobiology JournalPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713722957

Bacterial Calcium Carbonate Precipitation in Cave Environments: AFunction of Calcium HomeostasisEric D. Banksa; Nicholas M. Taylora; Jason Gulleya; Brad R. Lubbersa; Juan G. Giarrizoa; Heather A.Bullenb; Tori M. Hoehlerc; Hazel A. Bartona

a Department of Biological Sciences, Northern Kentucky University, Highland Heights, Kentucky b

Department of Chemistry, Northern Kentucky University, Highland Heights, Kentucky c NASA AmesResearch Center, Moffett Field, California

Online publication date: 08 June 2010

To cite this Article Banks, Eric D. , Taylor, Nicholas M. , Gulley, Jason , Lubbers, Brad R. , Giarrizo, Juan G. , Bullen,Heather A. , Hoehler, Tori M. and Barton, Hazel A.(2010) 'Bacterial Calcium Carbonate Precipitation in CaveEnvironments: A Function of Calcium Homeostasis', Geomicrobiology Journal, 27: 5, 444 — 454To link to this Article: DOI: 10.1080/01490450903485136URL: http://dx.doi.org/10.1080/01490450903485136

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

Geomicrobiology Journal, 27:444–454, 2010Copyright © Taylor & Francis Group, LLCISSN: 0149-0451 print / 1521-0529 onlineDOI: 10.1080/01490450903485136

Bacterial Calcium Carbonate Precipitation in CaveEnvironments: A Function of Calcium Homeostasis

Eric D. Banks,1 Nicholas M. Taylor,1 Jason Gulley,1 Brad R. Lubbers,1

Juan G. Giarrizo,1 Heather A. Bullen,2 Tori M. Hoehler,3 and Hazel A. Barton1

1Department of Biological Sciences, Northern Kentucky University, Highland Heights, Kentucky2Department of Chemistry, Northern Kentucky University, Highland Heights, Kentucky3NASA Ames Research Center, Moffett Field, California

To determine if microbial species play an active role in the de-velopment of calcium carbonate (CaCO3) deposits (speleothems) incave environments, we isolated 51 culturable bacteria from a coral-loid speleothem and tested their ability to dissolve and precipitateCaCO3. The majority of these isolates could precipitate CaCO3

minerals; scanning electron microscopy and X-ray diffractrometrydemonstrated that aragonite, calcite and vaterite were produced inthis process. Due to the inability of dead cells to precipitate theseminerals, this suggested that calcification requires metabolic ac-tivity. Given growth of these species on calcium acetate, but thetoxicity of Ca2+ ions to bacteria, we created a loss-of-function geneknock-out in the Ca2+ ion efflux protein ChaA. The loss of this pro-tein inhibited growth on media containing calcium, suggesting thatthe need to remove Ca2+ ions from the cell may drive calcification.With no carbonate in the media used in the calcification studies,we used stable isotope probing with C13O2 to determine whetheratmospheric CO2 could be the source of these ions. The resultantcrystals were significantly enriched in this heavy isotope, suggest-ing that extracellular CO2 does indeed contribute to the mineralstructure. The physiological adaptation of removing toxic Ca2+ ionsby calcification, while useful in numerous environments, would beparticularly beneficial to bacteria in Ca2+-rich cave environments.Such activity may also create the initial crystal nucleation sites thatcontribute to the formation of secondary CaCO3 deposits withincaves.

Received 25 May 2009; accepted 10 November 2009.The authors would like to thank the landowners and cavers in the

collection of the coralloid samples and strains, Dr. Dave Bunnell forthe image used in Figure 1, Dr. John Roth and Dr. Eric Kofoid forthe Salmonella strains and Mr. Michael D. Kubo for his assistancewith the isotopic analyses and IRMS work. EDB was supported by aSURF Fellowship from the ASM and a SURCA Award from NKU.HAB is supported in part by the Kentucky NSF EPSCoR Program(NSF0814194) and an NSF MIP/CAREER grant (NSF0643462), withinfrastructure support by the NIH Kentucky INBRE program (NIH5P20RR01648-05) and NSF Major Research Instrumentation award(MRI-0520921).

Address correspondence to Hazel A. Barton, Department of Bio-logical Sciences, Northern Kentucky University, SC 204D Nunn Drive,Highland Heights, KY 41099. E-mail: bartonh@nku.edu

Keywords calcite, calcium caves, coralloids, homeostasis,speleothems

INTRODUCTIONWhen secondary cave deposits (speleothems, a.k.a. cave for-

mations) were described in 1676 by John Beaumont, he classi-fied them as a form of plant life with “. . . growth from the finestparts of clay, being commonly white” (Beaumont 1676; Hilland Forti 1997). The description of cave formations as a vege-tative growth seems quaint as we now know that such depositsform when surface water percolating through the soil acquiresdissolved CO2, making a weak carbonic acid (H2CO3) that dis-solves limestone (calcium carbonate: CaCO3) rock. When thiswater drips into a cave, the CO2 off-gases and the subsequentincrease in the saturation index (SI) for Ca2+ and CO2−

3 leadsto the precipitation of CaCO3 as calcite, albeit at a exceedinglyslow rate (Hill and Forti 1997; Short et al. 2005b). Over ge-ologic timescales, this deposition leads to the development ofclassic speleothems, such as stalactites and stalagmites, as wellas a myriad of other potential forms (Hill and Forti 1997). Theinorganic and physical chemistry that drives these phenomenaare well understood, particularly from the works of Dreybrodtand Kaufmann et al. (Buhmann and Dreybrodt 1984; Dreybrodt1999; Kaufmann 2003; Short et al. 2005a, 2005b).

It has been known since the 1970’s that CaCO3 deposition(calcification) is a general phenomenon among the Bacteria(Boquet et al. 1973), with a high percentage of soil bacteriaproducing CaCO3 crystals when grown on media containingcalcium acetate. Despite the broad taxonomic distribution ofthis phenotype, much of the work examining bacterially me-diated CaCO3 precipitation has concentrated on cyanobacteriain Ca2+-rich (∼40 mm) seawater. In these environments, pho-tosynthesis alters the chemistry of the microenvironment byfixing CO2 (Aizawa and Miyachi 1986; Badger and Price 1994;Banfield and Nealson 1997), leading to an increase in the lo-cal pH. This favors the formation of CO2−

3 from HCO−3 and

leads to calcification. Within cave environments, which lack

444

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

BACTERIAL PRECIPITATION OF CALCIUM CARBONATE 445

photosynthetic activity and are found predominantly withinlimestone (CaCO3) rock, bacterial species demonstrate a higherincidence of the ability to precipitate calcium carbonate and anoverall increase in total CaCO3 precipitation levels (Cacchioet al. 2004; Danielli and Edington 1983).

Although bacterial fossils are found within speleothems(Melim et al. 2008; Melim et al. 2001), the observation of vi-able cells with these formations tends to be anecdotal (Baskaret al. 2006a; Cacchio et al. 2004; Danielli and Edington 1983);no cause-and-effect for the presence of microbial species inspeleothems has been elucidated (Barton and Northup 2007).If microbial species do play a role in the development ofspeleothems, it is difficult to understand the evolutionary ad-vantage of maintaining such a phenotype; when microorganismsbecome covered in an impermeable CaCO3 layer, this preventssolute exchange with the environment and ultimately leads todeath.

Among speleothems, coralloids are one of the most com-monly observed formations (Figure 1A) (Hill and Forti 1997).Various mechanisms for their formation have been described,from splashing of CaCO3 saturated water to the molecular move-ment of such water under capillary action (Hill and Forti 1997).Among these proposed methods, the need for surface irregu-larities on the base rock is consistent. Such irregularities createthe nucleation site for incorporation of additional crystal ele-ments, allowing subsequent CaCO3 deposition and speleothem

development (Banfield and Nealson 1997; Hill and Forti 1997).Through our microbiology studies in caves, we commonly iden-tify such ‘surface irregularities’ as microbial colonies; SEMexamination of these colonies often reveals CaCO3 crystal de-position on the surface of the bacterial cells (Barton and Northup2007).

These calcified colonies are often within close proximity tosmall CaCO3 mineral protrusions (∼1–2 mm), which enlargeup to the size of coralloids (Figure 1A). The close associationbetween calcified microbial colonies and coralloids suggestedthat microbial cells were either being caught up in the coralloidformation processes (Baskar et al. 2006b), or were somehowcritical in the development of these speleothems. The goal ofthis study was therefore to test whether calcification by bacterialspecies may be an important phenotype for survival in highcalcium cave environments, and whether such activity may leadto speleothem development.

MATERIALS AND METHODS

Sample SiteSamples were collected from an unnamed cave in Kentucky,

USA (cave conservation practices restrict naming the cave, butlocation and access information can be obtained from the au-thors). This ∼7 mile long cave is formed in Mississippian car-bonate rocks of the St. Genevieve and St. Louis Formation.

FIG. 1. (A) Coralloids (cave popcorn) is a common speleothem found in caves, for which no clear mechanism of formation has been described. (B) An examplethin-section through cave popcorn under polarizing light. Note the banding commonly observed in stromatolites. (C) and (D) The same thin-section image as B,but using EDS analysis. Colors correspond the element being mapped. Image A is ∼10 cm across.

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

446 E. D. BANKS ET AL.

Samples were collected in a stream passage above the caveflood line.

Bacterial CultivationIn order to culture bacterial isolates, swabs were wetted in

filtered (0.2 µm) cave water and developing coralloids wereswabbed. The swabs were then used to inoculate B-4 (Boquetet al. 1973) media and B-4C media (this study) containingB-4 media with a 1.5% top agar with 2% calcium carbonate(CaCO3) light powder (all chemicals, unless otherwise stated,were from Sigma Aldrich Chemicals, St. Louis, MO). Colonieswere picked and purified based on observation of CaCO3 pre-cipitation (P class) and/or dissolution (D class) and purified bypassage by Tryptic Soy Agar (TSA; Becton, Dickinson, Co.,Franklin Lakes, NJ). To compare the phenotype of non-cave de-rived isolates, additional species were obtained from the ATCC(Manassas, VA), including Microbacterium esteraromaticum(ATCC 51822), Stenotrophomonas maltophilia (ATCC 17666),Streptomyces senoensis (ATCC 29794) and Raoultella planti-cola (ATCC 33531) to supplement the lab strains Salmonellaentrica subsp. Typhimurium LT2 and Escherichia coli K12.

Mineral AnalysisIn order to examine the mineral being precipitated, CaCO3

precipitating isolates were grown in 3 ml of liquid B-4 me-dia. CaCO3 crystals were collected by filtration on 5.0 µm fil-ters (Milipore, Billerica, MA), washed free of cells with filter-sterilized distilled water (pH 8.0), air dried for thirty minutes,sputter coated with Au/PD and analyzed using an FEI Quanta200 environmental SEM. To confirm mineral identity, CaCO3

crystals were similarly collected for XRD analysis on a RigakuUltima III powder XRD using Cu Kα-radiation (40 kV, 44 mA).Data was collected over a 2θ range between 25–70◦ at scan rateof 0.1◦/min.

For isotopic probing to determine the source of the carbonatein the precipitated minerals, 3 ml liquid B-4 media was inocu-lated with the test strain and grown in a 2.4 liter glass chamber(Diagnostic Systems, Sparks, MD) fitted with an air tight gasket.C13O2 gas was then injected to generate an atmosphere that con-tained 0.2% atmospheric CO2 and 0.2% C13O2. Cultures weremaintained in this chamber for 2 weeks. Isotope ratio massspectrometry was carried out on a Nuclide 6-60 RMS dual-inletIRMS. To liberate the CO2 from the crystals, a degassed acidsolution was prepared. Briefly, 100 µl concentrated (∼10 M)phosphoric acid was added to 5 ml ultraclean (MilliQ) water.This acidified solution was degassed by repeated freeze/thawevents under vacuum, followed by 2 hours under vacuum.The acidified solution was then kept frozen on a methanol/dryice slurry, the sample container was removed from vacuum, thesample crystals were placed on the side of the reaction ves-sel and the vessel was put back under vacuum to evacuate theheadspace for an additional hour. The vessel was then sealed,taken off of vacuum, and the ice slurry was then allowed to melt

and flow down onto the crystals, producing a ‘fizzing’ that in-dicated the liberation of CO2. Samples were then run on a CO2

vacuum extraction line to isolate the liberated gas and trappedin glass ampoules for analysis against known standards.

GeologyTo examine the physical structure of the coralloids, thin sec-

tions were prepared from three rock samples collected usinga hammer and chisel, dried at 100◦C and impregnated withPetropoxy (Burnham Petrographics, Rathdrum, ID) while hot.The thin sections were prepared and photographed under planepolarized light using a Nikon E400Pol polarizing microscope.The geochemistry of the thin sections was analyzed by EDSmapping, as described in Spear et al. (2007).

Molecular TechniquesGenomic DNA was extracted from each isolate in

order to identify precipitating bacterial strains using aZR Fungal/Bacterial DNA Kit (Zymo Research, Orange,CA). The 16S ribosomal RNA gene sequence was PCRamplified as previously described (Spear et al. 2007)using the 8F (5′-AGAGTTTGATCMTGGCTCAG-3′) and1525R (5′-AAGGAGGTGATCCAGCC-3′) primers. Sequenc-ing of purified PCR products using the primers 8F and1391R (5′-GACGGGCGGTGWGTRCA-3′), with 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) as an internal primer wascarried out commercially by Agencourt Bioscience (Beverly,MA). All sequences were deposited in the NCBI database underaccession numbers FJ347997 – FJ348046. To identify isolates,gene sequences were compared to the Greengenes database(http://greengenes.lbl.gov) and confirmed by 16S rRNA phylo-genetic placement; DNA alignments were carried out using theARB Software Package (http://mpi-bremen.de/molecol/arb).

For distance calculations, the evolutionary model was GTR +I + G (base (0.2402, 0.2366 0.3148), Nst/6, Rmat (0.7547,1.8102, 1.0282, 0.7316, 3.1058), Rates = gamma, Shape =0.6293, Pinvar = 0.2486), determined using MODELTEST(Posada 2003). Distance and parsimonious phylograms weregenerated in PAUP* using a heuristic search (Wilgenbuschand Swofford 2003). The highest scoring trees were tested byboostrap analysis using 1000 replicates. A maximum likeli-hood phylogram was also generated using a Bayesian analysiswith the MrBayes program for 2 million generations (Ronquistand Huelsenbeck 2003). In all cases, sequences from Aquifexpyrophilus and Thermoplasma acidophilium were used foroutgroups.

Construction of Deletion MutantsTo examine the role that microbial physiology had in precip-

itation, deletion mutants of the genes chaA and nhaA were gen-erated using linear transformation in S. typhimurium TT22971containing a lambda red background, based on the protocol ofDatsenko and Wanner (2000). To confirm that the insert had

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

BACTERIAL PRECIPITATION OF CALCIUM CARBONATE 447

TABLE 1Primers used in this study

Primer Name Sequence

yadF-forward 5′-ATTGGTGGGGAATGTATTTCAGACTTAGGGCGGAAATACCACCA AACACCCCCCAAAACC-3′

yadF-reverse 5′-CCCAACGCACTTATTTGGTTACAGGTCGTTAACTTCCATCACACAACCACACCACACCAC-3′

yadF-FPE 5′-TTGGCTTTTCGATTCCGGACG-3′

yadF-RPI 5′-ATCTGAACTGTCTCTCCGTGG-3′

yarF-RPE 5′-ATTTCCCTGCTCCATTTGGC-3′

chaA-forward 5′-ATGACACATGCCTGTGAGGCGGTGAAAACCCGCCATAAGCACCAAACACCCCCCAAAACC-3′

chaA-Reverse 5′-TGCGGTTTTTAGCTTGTGTAGGCCGGATAAGATACTATCCACACAACCACACCACACCAC-3′

chaA-FPE 5′-CCCAGTCGAGGAGGATTT-3′

chaA-RPI 5′-GACCAATAATGCCTGACCG-3′

chaA-RPE 5′-AATGCCGCGACCCATTTG-3′

knocked out the gene being studied, PCR was carried out usinga combination of a forward external primer (FPE) and either areverse internal (RPI) or external (RPE) primer (Table 1). Onestrain that contained either a deletion of chaA or nhaA was thentransduced by P22 into a wild-type S. typhimurium TR10K back-ground and verified by PCR to generate the respective knockout mutant.

RESULTS

Analysis of Coralloid Thin SectionsEight thin sections were prepared from rock samples that con-

tained both observed microbial colonies and 56 small, nodularcoralloids (an example is shown in Figure 1B–D). Within thesethin sections, polarizing-light microscopy revealed a dark, lam-inated layer between the base of each coralloid and the hostrock (Figure 1B), which matched the thin-section appearancethrough colonies that had been visually identified on the surfaceof these samples. In all cases, this dark material was coated withlayers of what appeared to be CaCO3 (by gross examination)in a stromatolitic-like growth, quite distinct from the mineralstructure of the host rock (as shown in Figure 1B). To examinethe chemistry of this deposition, energy-dispersive spectroscopy(EDS) mapping of the thin-sections was carried out (example inFigures 1C and 1D).

The results demonstrate that the coralloids were calcium-and oxygen-rich in chemistry (indicative of CaCO3), lackingthe trace silica, aluminum and iron signatures of clays ob-served in the St. Genevieve Formation host rock (Figures 1B–D)(McGrain and Dever 1967). These data suggest that the struc-ture and chemistry of these coralloids is different than the hostrock and in-line with a depositional origin. There was also whatappeared to be an accumulation of silica, aluminum and iron-rich material between all examined coralloids and the host rock(Figure 1C and 1D). The chemistry of this material matchesthe chemistry of clay minerals in the Ste. Genevieve Formation,which are generally not soluble, except under extremely low pH.

Distribution of the Calcification Phenotype AmongBacteria Isolated From Speleothems

To assess whether microbial activity played a role in theCaCO3 deposition seen in coralloid development, bacteria wereisolated from coralloids in close proximity to the rock samples.This was done by streaking swab samples onto either B-4 mediaor B-4C media, which allow us to identify bacterial species thatdemonstrate the ability to either precipitate or dissolve CaCO3,respectively (Figure 2). Following growth at room temperaturefor 2–4 weeks, isolates were identified based on unique colonymorphology and phenotype; species able to precipitate CaCO3

were considered P-class, and those able to dissolve calcite wereD-class isolates. This initial screen on B-4 and B-4C mediaidentified 15 P-class and 17 D-class colonies. These colonieswere then picked and streaked for pure culture on TSA media,which often allows the separation of mixed cultures. Using thisapproach we identified 19 additional (10 P-class and 9 D-class)isolates, generating 51 unique bacterial cultures isolated fromthese developing coralloids.

To ensure that the crystals produced by these isolates onB-4 media were indeed CaCO3 mineral polymorphs, they wereexamined using SEM microscopy and X-ray diffraction (XRD).

The SEM results (Figure 3) demonstrated the presence ofmineral polymorphs, including calcite, vaterite and aragonite(Lippmann 1973). XRD analysis confirmed the presence ofthese multiple mineral forms, including vaterite, which is onlytransiently found in nature (Figure 4). Although examining thesecrystals using SEM, it was noted that virtually all of the crys-tals included pits and depressions that matched the structure ofthe precipitating bacteria, both in terms of size and morphol-ogy (Figures 3D and 3E). A number of these depressions alsoincluded structures that were indicative of cell growth and di-vision, such as the septa observed in Figure 3E. Even thoughthese crystals were washed free of bacterial cells, many crystalsalso contained obvious cells closely associated with, and emerg-ing from, the mineral surface (Figure 3F). The organic nature ofthese cells was confirmed by EDS, which indicated the presence

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

448 E. D. BANKS ET AL.

FIG. 2. Precipitation and dissolution phenotypes of cave isolates. (A) Original colony morphologies of isolates streaked on B-4 media. The calcium carbonateprecipitates demonstrating different mineral polymporphs can be seen within each colony. Colonies range in size from 2–4 mm. (B) Calcite dissolution by D17Dwhen grown on B-4C media. The zone of clearing of the CaCO3 around each colony can be seen, along with CaCO3 precipitation in the center of isolated colonies.

of phosphorous, nitrogen and sulfur peaks when compared tothe calcium carbonate mineral (data not shown). Together thesedata demonstrate a close association between the bacterial cellsand the formed crystals.

Once calcification by the bacterial isolates was confirmed,PCR amplification of the 16S rRNA gene sequence was usedto determine their identity. The results (Figure 5) demonstratea broad distribution of isolates, representing members of the

Alpha-, Beta- and Gammaproteobacteria, the Firmicutes andActinobacteria. While this distribution does not represent the to-tal bacterial population in this environment, the divisional repre-sentation is in line with previous cultivation and non-cultivationbased studies in cave and subsurface environments (Amann etal. 1995; Barns and Nierzwicki-Bauer 1997; Barton et al. 2004).With the exception of GGC-D3, all 51 isolates demonstrated acalcification phenotype on B-4 media, irrespective of whether

FIG. 3. SEM images of calcite crystals generated by GGC cultivars. Numerous calcite polymorphs are produced by microbial activity, including (A) calcite(P9B), (B) aragonite (D15) and (C) vaterite (P11). The close associate between microbial growth and the precipitation of these minerals can also be seen, includingcell-shaped pits (D) and septa (E); indicated by arrow). Although these crystals were washed free of bacterial cells prior to SEM, the close association between thebacteria and the crystals can be seen by the emergence from and continued adherence of cells to the crystals (F).

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

BACTERIAL PRECIPITATION OF CALCIUM CARBONATE 449

FIG. 4. Representative XRD analysis of calcium carbonate crystals (fromisolate P11) shows presence of vaterite. The corresponding peaks are labeled.

they were initially identified based on a phenotype of precipi-tating or dissolving calcium carbonate (Figure 5).

Dissolution activity was not so broadly represented, likelydue to the large amount of acid production necessary to ob-serve this phenotype; however, many of the D-class isolatesalso precipitated CaCO3 crystals on top of their colonies dur-ing dissolution (Figure 5). This phenotype was observed for alldissolving phenotypes, except GGC-D10A, GGC-D10B1 andGGC-P11, which generated a very large zone of clearing.

The pattern of both CaCO3 precipitation, dissolution anda combined dissolving/precipitating phenotype appeared to besimilar for all isolates within a particular clade, with the ex-ception of the Bacilli (Figure 5). To determine whether therelationship between members of a clade and phenotype is suf-ficiently robust to be predictive of phenotype, five bacterialspecies isolated from non-cave environments were tested fortheir ability to precipitate and/or dissolve CaCO3; these includedS. typhimurium, Escherichia coli, Klebsiella planticola, Mi-crobacterium esteraromaticum, Stenotrophomonas maltophiliaand Streptomyces senoensis. The results (Figure 5) demonstratethat the phenotype for these species does indeed match those forthe clade to which they belong, despite the fact that they werenot isolated from caves.

The Physiology of Calcium Carbonate PrecipitationThe close association between clade and phenotype shown

in Figure 5 suggests that either a structural or genotypic re-lationship is responsible for the CaCO3 phenotypes observed.To differentiate the contribution of bacterial structure to crys-tal formation, the precipitation capabilities of a random groupof isolates killed with paraformaldehyde was tested; the lossof cell viability using this method was confirmed by plating.Killed cells were unable to precipitate CaCO3 crystals in liquid

B-4 media, suggesting that cells must be metabolically activefor calcification and that cell structure alone is not sufficientto promote this process. In support of a genetic basis, it wasnoted that during the course of these experiments, the CaCO3

precipitation phenotype for many isolates on B-4 was lost whenthey were cultivated on TSA media. Yet, calcification on B-4media could be restored by prior passage of the culture on mediacontaining either calcium acetate or CaCO3 powder.

To assess whether Ca2+ ions were providing the selectionpressure for maintenance of the calcification phenotype, thegene for the Ca2+/2H+ antiporter chaA (Ivey et al. 1993) wasknocked out in S. typhimurium LT2, removing its function withinthe cell. The gene nhaA (which encodes a Na+/H+ antiporter)was also knocked out as a control against the effects of a geneknockout on the general precipitation phenotype (Ohyama et al.1994). In order to assess the role that these mutations had oncalcium carbonate precipitation, wild type (WT), �chaA and�nhaA strains were streaked on B4-media and a 0.1% glu-cose/0.5% yeast extract media amended with 4g/l CaCO3. Whilethe �chaA and �nhaA mutants grew as well as WT on TSAmedia, their growth was inhibited on media containing calcium(Figure 6). The �chaA mutant also grew poorly on B-4 mediawhen compared to wild type. While this mutant did grow weaklyon the glucose/yeast extract media amended with CaCO3, themedia could not be cleared of this mineral (Figure 6) whencompared to WT. The �nhaA control did not demonstrate asignificant phenotype under any of the conditions used.

If Ca2+ ions are transported out of the cell by chaA duringgrowth on calcium-rich substrates, the source of CO2−

3 ions forsubsequent CaCO3 precipitation remained unclear; B4 mediadoes not contain CO2−

3 ions. We therefore tested whether at-mospheric CO2 played a role in crystal formation using stableisotope probing. Cave isolates D7A, D11 and D9A were grownin liquid B-4 media under normal atmospheric air (control) oran atmosphere amended with 0.2% C13O2. These cultures wereincubated for 2 weeks at room temperature and the crystals thatformed were harvested and tested for the presence of heaviercarbon isotopes using IRMS.

The results for the control sample (D7A), in which the at-mosphere was not supplemented with C13O2, were in line withspeleothem studies, with a slight enrichment for the lighter iso-tope (δ13Cvpdb = −12.53‰ ± 0.10‰). In comparison, the crys-tals of D11 and D9A grown in the presence of a C13O2 werehighly enriched for the heavier isotope. D11 was measured atδ13Cvpdb + 403.3‰ ± 0.12‰, while D9A was too enriched inC13 to be accurately measured.

DISCUSSIONIn studying cave microorganisms, we have long questioned

the mechanisms that allow bacterial species to survive underthe very high levels of calcium encountered in these environ-ments (Barton et al. 2001, 2004, 2007). It is known that micro-bial species from cave environments demonstrate an enhanced

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

450 E. D. BANKS ET AL.

FIG. 5. Phylogenetic placement of GGC cultivars with their corresponding calcite precipitation and dissolution phenotypes: � = strong phenotype; =intermediate phenotype; � = absent. The ∗ symbol indicates that crystals were also precipitated during dissolution, while indicates a very high level of CaCO3

dissolution around the colony. The phenotypes from a number of previously characterized bacterial strains are also included (in green). The phylogram representsa consensus tree generated from distance and maximum likelihood tree-building algorithms. The scale bar represents 0.1 substitutions per site.

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

BACTERIAL PRECIPITATION OF CALCIUM CARBONATE 451

FIG. 6. Phenotypes of knockout mutants grown on B-4 media or glucose/yeastmedia containing CaCO3. Each of the mutants (�chaA and �nhaA) is comparedto wild-type (WT). The dark zones seen on the CaCO3 media represents clearingof this mineral from the media, allowing the dark background to be observed.

capability to precipitate calcium carbonate minerals, both fromthe total number of species displaying this phenotype to theextent they can precipitate this mineral (Danielli and Edington1983). The survival of these species under such conditions iseven more remarkable given the predominance of heterotrophicmicrobial species in caves, the metabolic activity of which maylead to the formation of organic acids. If such acids are produced,these can dissolve the host rock (CaCO3) as shown in Figure 2.

If microbial activity in caves is dissolving the host rock,this would generate the calcium salt of that acid. In support ofthis hypothesis, we have previously observed the in situ pro-duction of both calcium acetate and calcium pyruvate in caveenvironments (Bullen et al. 2008). This observation is signifi-cant, given that calcium acetate has been the staple of assessingthe bacterial calcification phenotype since the 1970s (Boquetet al. 1973). In this study, we suggest that such dissolution ac-tivity is further evidenced by the accumulation of insoluble clayresidues underneath apparent microbial colonies on the host rock(Figure 1C–D). While such clay deposits are also formed duringthe process of speleogenesis, they were not detected elsewhereon the samples and only directly below the apparent microbialcolonies. Similar clay accumulation under speleothem depositshas been observed in other caves (Barton and Northup 2007;Blyth and Frisia 2008; Borsato et al. 2000).

Past investigators have argued that passive structural com-ponents of the bacterial cell play the most important role incalcification (Barabesi et al. 2007; Wolfe 2005). Nonetheless,many of these studies were carried out in Bacilli species, whichcontain numerous mineral precipitating polymeric surface struc-tures (Ercole et al. 2007; Sleytr and Beveridge 1999) and havethe least consistent precipitation phenotype (Figure 5). Suchstudies have argued against a metabolic role in calcification, askilled cells also promote CaCO3 precipitation (Martinez et al.2008; Yates and Robbins 1998); however these studies killedcells by autoclaving or using metabolic poisons, such as potas-

sium cyanide and sodium azide. It is hard to determine what con-sequence autoclaving would have on the chemistry of CaCO3

precipitation as it leads to cell lysis and gross structural changes.The ability of metabolic poisons to kill bacterial cells can be lim-ited by cyanide/azide resistant cytochrome oxidases (Cunning-ham and Williams 1995; Lichstein and Soule 1943) or the abilityof fermentation reactions to maintain cellular viability. Indeed,we found that a large number of our isolates remained viablefollowing similar cyanide/azide treatments (data not shown). Inthis study treatment with paraformaldehyde, which maintainscellular structure by cross-linking proteins, appeared to consis-tently kill the cells. This approach may explain why our resultswere more consistent in ruling out a role for metabolic activity.

This work, in support of other investigators (Buczynski andChafetz 1991), therefore suggests that bacterial metabolismplays a dominant role in calficifaction. In attempting to deducea metabolic role for calcification where entombment would beultimately detrimental to survival, it is important to examine thecontext of the environment in which the cell is found. Underthe nutrient limitation of cave environments, any reduced com-pounds excreted by he cell are likely to be reassimilated as anenergy source, as suggested by the presence of high-affinity ac-etate transporters encoded by a large number of bacterial species(Wolfe 2005). Indeed, in a past study, we saw the assimilationof a number of metabolic products during the growth of E. coliin the presence of calcite crystals (Bullen et al. 2008).

However, species that uptake these potential carbon and en-ergy sources in calcium-rich environments must also deal withtoxic Ca2+ ions (Anderson et al. 1992). The ChaA antiporter,which is conserved across many bacterial phyla (Ivey et al.1993), can secrete these ions into the extracellular environment,but must do so against an ever-increasing concentration gradient(Anderson et al. 1992).

An alternative approach for these cells could be to remove ex-cess Ca2+ ions as they are excreted by the cell, via calcification.The high distribution of this phenotype among these isolatesin accordance with the results of previous speleothem cultiva-tion studies (Danielli and Edington 1983). Indeed, the majorityof the CaCO3 dissolving isolates also re-precipitated CaCO3

on the surface of the colony during growth (Figures 2 and 5).The only exceptions to this were GGC-D10A, GGC-D10B1 andGGC-P11; however, these species demonstrated a remarkableamount of CaCO3 dissolution and presumably a very high levelof acid production by these strains is not compatible with CaCO3

co-precipitation. This could also explain the widespread CaCO3

precipitation phenotype among bacteria as Ca2+ ions are foundin a wide variety of habitats.

The work of Barabesi et al. (2007) has recently identifieda potential role for fatty acid metabolism in CaCO3 precipi-tation by B. subtilis. A mutation in fadD, which regulates theβ-oxidation of fatty acids to acetyl-CoA, prevented calcifica-tion on B-4 media. While the exact role of this gene in CaCO3

precipitation remains to be elucidated, these investigators car-ried out experiments on B-4 media containing calcium acetate.

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

452 E. D. BANKS ET AL.

FIG. 7. The glyoxylate cycle (also known as the glyoxylate bypass or gly-oxylate shunt) used in the consumption of acetate for central metabolism. Inaddition to energy production, acetate is also used for fatty acid synthesis, whileoxaloacetate is consumed for gluconeogenesis. The putative roles of yadA andchaA are also shown.

Using calcium acetate as a carbon and energy source for growthpresents a problem for bacterial cells due to the assimilationof acetate; to allow conservation of CO2 for gluconeogenesis,acetate must enter the TCA cycle via the glyoxylic acid bypass,limiting the availability of CO2 for cellular processes such asfatty acid synthesis (Figure 7).

The production of acetyl-CoA from fatty acid catabolismby FadD also requires the cell to use the glyoxylate bypassfor growth (Figure 7), affecting the overall CO2 budget forthe cell. Given that our results suggest that atmospheric CO2

is at least partially responsible for the CO2−3 ions found in the

calcium carbonate crystals, it is unclear how altering the cellularCO2 budget may affect this process. Calcification experimentshave also been carried out in ureolytic bacterial species, whichbreakdown urea with the generation of ammonia (Hammes et al.2003). This activity leads to a sharp increase in pH, promotingcalcite precipitation; however, this process is primarily gearedtoward the solid-phase capture of organic contaminants andthe high levels of urea used in these experiments (approaching600 mm) are not normally found in nature (Hammes et al. 2003).

The chaA gene knock-out generated in this study and the ob-served phenotype on calcium containing media provides the firstevidence for a link between calcium metabolism in bacteria andcalcification. Although the extrusion of Ca2+ ions would cer-tainly contribute to bacterial precipitation of CaCO3, the sourceof CO2+

3 ions is more difficult to discern, especially given thatthe carbonic anhydrase gene (yadF) appears to be an essentialgene for a large number of bacterial species under normal atmo-spheric conditions (Kusian et al. 2002; Merlin et al. 2003). Ourstable isotope probing analyses suggest that atmospheric CO2

is contributing, in part, CO2−3 ions to the precipitated minerals.

This may occur through the equilibrium of CO2 between theatmosphere and as bicarbonate ions in the medium surroundingthe cell (Formula 1).

pKa = 6.35 pKa = 10.33H2O + CO2 ↔ HCO−

3 + H+ ↔ CO2−3 + 2H+

[1]

The fate of the protons would influence subsequent carbonateion formation; however, it is possible that they would be con-sumed as part of cellular metabolism (contributing to the protonmotive force). Such activity would draw Equation 1 to the rightand favor the formation of carbonate ions. Indeed, past investiga-tors have argued that calcification generates protons for nutrientacquisition, with the phenotype becoming more pronounced un-der nutrient limitation (McConnaughey and Whelan 1997). Thiswould also correlate well with the extreme nutrient limitationof cave environments. Alternatively, these protons could be uti-lized by the ChaA Ca2+/2H+ antiporter, which would aid in thetransport of Ca2+ ions into the extracellular medium. Together,the loss of the protons would generate more alkali conditionsand favor the formation of CO2−

3 ions. When the concentrationsof both Ca2+ and CO2−

3 ions exceed the solubility product forthese ions (Ksp = 4.5 × 10−9 for calcite), calcification occurs.

The role of microbial species in the development of sec-ondary, carbonate deposits in caves has remained controversialfor quite some time (Barton et al. 2001), with mostly anecdotalevidence of microbial association within speleothems (Baskaret al. 2006a) and the lack of a cause-and-effect rationale (Bartonand Northup 2007). Whatever the pathway, bacterial metabolicactivity in these environments appears to lead to the precipita-tion of various CaCO3 mineral forms on the microbial colony.In the case of coralloids, these bacterially deposited crystals canthen form the nucleation site for subsequent mineral accumula-tion through abiotic processes (Banfield and Nealson 1997). Inthe formation of coralloids, such secondary processes may in-clude the moisture-driven flow of CaCO3 saturated water fromareas of dissolution to precipitation in caves containing con-vention airflow currents (M. Queen, personal communication,2007). Such activity would generate the distinct lines of coral-loid formation often seen in caves with such airflow patterns.Our results therefore suggest that, while the growth of coralloids

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

BACTERIAL PRECIPITATION OF CALCIUM CARBONATE 453

may be a physiochemical phenomenon, the precise location ofthese deposits may be influenced by the need of extant microbialspecies to maintain calcium homeostasis.

CONCLUSIONSThe bacterial precipitation of calcium carbonate minerals has

been dismissed by some investigators as a passive process, giventhe lack of material to link calcium metabolism with microbialecosystem energetics. Here, by examining the calcification phe-notypes of numerous bacterial species and using a knock-outin the chaA calcium antiporter protein, we suggest that calciumtoxicity provides both the physiological basis and selection pres-sure for the calcification phenotype. We propose that Ca2+ ionsare detoxified by calcification, which allows the cell to survivean immediate metabolic insult. While this approach might ul-timately prove fatal, from cellular entombment and death, thewide distribution of this phenotype suggests that the short-termbenefits may outweigh the potential cost. If the pathway we haveproposed is correct, then this could have a profound impact onour understanding of both microbially derived calcium precip-itation in a variety of habitats and the potential for long-term,geologic sequestration of atmospheric CO2.

REFERENCESAizawa K, Miyachi S. 1986. Carbonic anhydrase and CO2 concentrating mech-

anisms in microalgae and cyanobacteria. FEMS Microbiol Lett 39:215–233.Amann RI, Ludwig W, Schleifer KH. 1995. Phylogenetic identification and

in situ detection of individual microbial cells without cultivation. MicrobiolRev 59:143–169.

Anderson S, Appanna VD, Huang J, Viswanatha T. 1992. A novel role for calcitein calcium homeostasis. FEBS Lett 308:94–96.

Badger MR, Price GD. 1994. The Role of Carbonic Anhydrase in Photosynthe-sis. Ann Rev Plant Physiol Plant Mol Biol 45:369–392.

Banfield JF, Nealson KH. 1997. Geomicrobiology: Interactions between mi-crobes and minerals. Washington, DC: Mineralogical Society of American.448 p.

Barabesi C, Galizzi A, Mastromei G, Rossi M, Tamburini E, Perito B. 2007.Bacillus subtilis gene cluster involved in calcium carbonate biomineralization.J Bacteriol 189:228–235.

Barns SM, Nierzwicki-Bauer SA. 1997. Microbial diversity in modern sub-surface, ocean, surface environments. In: Banfield JF, Nealson KH, editors.Geomicrobiology: Interactions between microbes and minerals. Washington,DC: Mineralogical society of America. P 35–79.

Barton HA, Northup DE. 2007. Geomicrobiology in Cave Environments: Past,Current and Future Prospectives. J Cave Karst Stud 69:163–178.

Barton HA, Spear JR, Pace NR. 2001. Microbial life in the underworld: Bio-genicity in secondary mineral formations. Geomicrobiology J 18:359–368.

Barton HA, Taylor MR, Pace NR. 2004. Molecular phylogenetic analysis of abacterial community in an oligotrophic cave environment. GeomicrobiologyJ 21:11–20.

Barton HA, Taylor NM, Kreate M, Springer AJ, Oehrle SA, Bertog JL. 2007.The impact of host rock geochemistry on bacterial community structure inoligotrophic cave environments. Int J Speleol 36:93–104.

Baskar S, Baskar R, Mauclaire L, McKenzie JA. 2006. Microbially inducedcalcite precipitation in culture experiments: Possible origin for stalactites inSahastradhara caves, Dehradun, India. Indian Curr Sci 90:58–64.

Beaumont J. 1676. Two letters concerning rock plants and their growth: Thesecond letter of June 17, 1676. Philos Trans Royal Soc 11:732–742.

Blyth AJ, Frisia S. 2008. Molecular evidence for bacterial mediation of cal-cite formation in cold high-altitude caves. Geomicrobiology J 25:101–111.

Boquet E, Boronate A, Ramos-Cormenzana A. 1973. Production of calcite(calcium carbonate) crystals by soil bacteria is a general phenomenon. Nature246:527–529.

Borsato A, Frisia S, Jones B, van der Borg K. 2000. Calcite moonmilk: Crystalmorphology and environment of formation in caves in the Italian Alps. JSedimen Res 70:1179–1190.

Buczynski C, Chafetz HS. 1991. Habit of bacterially induced precipitates ofcalcium carbonate and the influence of medium vicosity on mineralogy. J.Sedimen Pet 61:226–233.

Buhmann D, Dreybrodt W. 1984. The kinetics of calcite dissolution and precip-itation in geologically relevant situations of karst areas 1 Open system. ChemGeol 48:189.

Bullen HA, Oehrle SA, Bennett AF, Taylor NM, Barton HA. 2008. Use ofattenuated total reflectance fourier transform infrared spectroscopy to identifymicrobial metabolic products on carbonate mineral surfaces. Appl EnvironMicrobiol 74:4553–4559.

Cacchio P, Contento R, Ercole C, Cappuccio G, Martinez MP, Lepidi A. 2004.Involvement of microorganisms in the formation of carbonate speleothems inthe cervo cave (L’Aquila-Italy). Geomicrobiology J 21:497–509.

Cunningham L, Williams HD. 1995. Isolation and characterization of mutantsdefective in the cyanide-insensitive respiratory pathway of Pseudomonasaeruginosa. J Bacteriol 177:432–438.

Danielli HMC, Edington MA. 1983. Bacterial calcification in limestone caves.Geomicrobiology J 3:1–16.

Dreybrodt W. 1999. Chemical kinetics, speleothem growth and climate. Boreas28:347–356.

Ercole C, Cacchio P, Botta AL, Centi V, Lepidi A. 2007. Bacterially InducedMineralization of Calcium Carbonate: The Role of Exopolysaccharides andCapsular Polysaccharides. Microsc Microanal 13:42–50.

Hammes F, Boon N, de Villiers J, Verstraete W, Siciliano SD. 2003. Strain-specific ureolytic microbial calcium carbonate precipitation. App EnvironMicrobiol 69:4901–4909.

Hill CA, Forti P. 1997. Cave Minerals of the World. 2nd ed. Huntsville: NationalSpeleological Society 463 p.

Kaufmann G. 2003, Stalagmite growth and palaeo-climate: the numerical per-spective. Earth Planet Sci. Lett 214:251–266.

Ivey DM, Guffanti AA, Zemsky J, Pinner E, Karpel R, Padan E, SchuldinerS, Krulwich TA. 1993. Cloning and characterization of a putative Ca2+/H+antiporter gene from Escherichia coli upon functional complementation ofNa+/H+ antiporter-deficient strains by the overexpressed gene. J Biol Chem268:11296–11303.

Kaufmann G. 2003. Stalagmite growth and palaeoclimate: the numerical per-spective. Earth Planet Sci Lett 214:251–266.

Kusian B, Sultemeyer D, Bowien B. 2002. Carbonic anhydrase is essential forgrowth of Ralstonia eutropha at ambient CO2 concentrations. J Bacteriol184:5018–5026.

Lichstein HC, Soule MH. 1943. Studies of the effect of sodium azide on microbicgrowth and respirations: I The action of sodium azide on microbic growth. JBacteriol 47:221–230.

Lippmann F. 1973. Sedimentary carbonate minerals. New York:Springer-Verlag.228 p.

Martinez RE, Pokrovsky OS, Schott J, Oelkers EH. 2008. Surface charge andzeta-potential of metabolically active and dead cyanobacteria. J. Coll InterfSci 323:317–325.

McConnaughey TA, Whelan JF. 1997. Calcification generates protons for nu-trient and bicarbonate uptake. Earth Sci Rev 42:95–177.

McGrain P, Dever GR. 1967. Limestone resources in the Appalachian region ofKentucky. Kentucky Geological Survey, Ser 10. Bull 4:1–12.

Melim LA, Northup DE, Spilde MN, Jones B, Boston PJ, Bixby RJ. 2008.Reticulated filaments in cave pool speleothems: Microbe or mineral? J. CaveKarst Stud 70:135–141.

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010

454 E. D. BANKS ET AL.

Melim LA, Shinglman KM, Boston PJ, Northup DE, Spilde MN, Queen JM.2001. Evidence of microbial involvement in pool finger precipitation, HiddenCave, New Mexico. Geomicrobiol J 18:311–330.

Merlin C, Masters M, McAteer S, Coulson A. 2003. Why is car-bonic anhydrase essential to Escherichia coli? J Bacteriol 185:6415–6424.

Ohyama T, Igarashi K, Kobayashi H. 1994. Physiological role of the chaAgene in sodium and calcium circulations at a high pH in Escherichia coli. JBacteriol 176:4311–4315.

Posada D. 2003. Selecting models of evolution. In: Vandemme A, Salemi M,editors. The Phylogenetic Handbook. Cambridge, UK: Cambridge UniversityPress. P 256–282.

Ronquist F, Huelsenbeck J. 2003. Mr Bayes 3: Bayesian phylogenetic inferenceunder mixed models. Bioinformatics 19:1572–1574.

Short MB, Baygents JC, Beck JW, Stone DA, Toomey RS, Goldstein RE.2005a. Stalactite growth as a free-boundary problem: A geometric law andits platonic ideal. Phys Rev Lett 94:018511–4.

Short MB, Baygents JC, Goldstein RE. 2005b. Stalactite growth as a free-boundary problem. Phys Fluids 17:083101–12.

Sleytr UB, Beveridge TJ. 1999. Bacterial S-layers. Trends Microbiol 7:253–260.Spear JR, Barton HA, Roberts JK, Francis CA, Pace NR. 2007. Microbial

community biofabrics in a geothermal mine adit. Appl Environ Microbiol73:6172–6180.

Wilgenbusch JC, Swofford D. 2003. Inferring evolutionary trees with PAUP*.Curr Protoc Bioinformatics Chapter 6, Unit 6.4.

Wolfe AJ. 2005. The acetate switch. Microbiol Mol Biol Rev 69:12–50.Yates KK, Robbins LL. 1998. Production of carbonate sediments by a unicellular

green alga. Am Mineral 83:1503–1509.

Downloaded By: [Barton, Hazel] At: 05:02 10 June 2010