Prevalence of Ca²⁺-ATPase-Mediated Carbonate Dissolution among Cyanobacterial Euendoliths

Abstract

Recent physiological work has shown that the filamentous euendolithic cyanobacterium Mastigocoleus testarum (strain BC008) is able to bore into solid carbonates using Ca²⁺-ATPases to take up Ca²⁺ from the medium at the excavation front, promoting dissolution of CaCO₃ there. It is not known, however, if this is a widespread mechanism or, rather, a unique capability of this model strain. To test this, we undertook a survey of multispecies euendolithic microbial assemblages infesting natural carbonate substrates in marine coastal waters of the Caribbean, Mediterranean, South Pacific, and Sea of Cortez. Microscopic examination revealed the presence of complex assemblages of euendoliths, encompassing 3 out of the 5 major cyanobacterial orders. 16S rRNA gene clone libraries detected even greater diversity, particularly among the thin-filamentous forms, and allowed us to categorize the endoliths in our samples into 8 distinct phylogenetic clades. Using real-time Ca²⁺ imaging under a confocal laser scanning microscope, we could show that all communities displayed light-dependent formation of Ca²⁺-supersaturated zones in and around boreholes, a staple of actively boring phototrophs. In 3 out of 4 samples, boring activity was sensitive to at least one of two inhibitors of Ca²⁺-ATPase transporters (thapsigargin or tert-butylhydroquinone), indicating that the Ca²⁺-ATPase mechanism is widespread among cyanobacterial euendoliths but perhaps not universal. Function-community structure correlations point to one particular clade of baeocyte-forming euendoliths as the potential exception.

INTRODUCTION

Euendolithic cyanobacteria, also known as boring, microboring, excavating, perforating, or tunneling cyanobacteria, can penetrate a variety of calcareous substrates, such as shells, dead coral, and limestone, by means of chemical dissolution. Aside form cyanobacteria, only fungi and some eukaryotic algae are capable of such boring activity. A wide variety of euendolithic cyanobacteria has been described during almost 2 centuries of naturalistic studies (4, 9, 14, 23). While only a small proportion of these species are capable of boring, this select group has representatives in several of the major taxonomic groups of cyanobacteria. They can be found in diverse geographical locations worldwide (1, 2, 5, 24), where they are involved in the diagenesis of the substrates they colonize. In most instances, they weaken the structure of the solid substrate with their tunneling (26, 27), but in other cases, the reprecipitation of carbonates as micrite, a by-product of dissolution, forms the cement that binds carbonate sand grains into stromatolites (21).

The mechanism that allows cyanobacterial endoliths to bore, a subject of long-standing controversy (6, 11), has been recently elucidated in the model filamentous strain Mastigocoleus testarum BC008 (20). It is based on the uptake and transcellular transport of Ca²⁺ (12). This is driven, at least partly, by P-type ATPases. As it is presently understood, the trichomes import free Ca²⁺ from the interstitial space between the boring cell and mineral (“boring front”), lowering the concentration of the ion in the interstitial space below that of saturation and thus promoting the localized dissolution of the calcium carbonate at the boring front. Intracellular Ca²⁺ is transported from cell to cell along the filament, likely with the aid of Ca²⁺-specific pumps or channels, and excreted at the end farthest from the boring front. This ATP-powered extrusion results in very strong Ca²⁺ supersaturation around the boreholes. The mechanism, shown in M. testarum strain BC008, differs from the previously tacitly accepted hypothesis contending that boring is powered by mere secretion of acid. Because of the diversity of boring cyanobacteria in nature, it is only natural to question if the strategy of BC008 is universal. While a mechanism common to all cyanobacterial microborers seems plausible, it is difficult to evaluate a priori on the basis of phylogenetic relatedness, given that, in fact, explicit modern phylogenetic work on euendoliths is extremely restricted; references 5 and 8 provide the two sequences available.

The overarching goal of this work was to address the potential universality of the boring mechanism among cyanobacterial euendoliths, particularly given that the model strain cannot bore on dolomitic substrates (20). The boring mechanism was investigated empirically in complex microbial assemblages collected from distinct geographical locations worldwide, so as to encompass as diverse as possible a range of subjects, and compared to the one recently described for M. testarum BC008.

MATERIALS AND METHODS

Field samples.

Euendolith-infested biogenic carbonates (shells and dead coral skeleton pieces) were collected from four coastal regions, the Caribbean (San Juan, Puerto Rico, 18°45′N, 65°96′W), North Pacific Sea of Cortez (Baja California, 30° 92′N, 114°70′W), Mediterranean (L'Alguer, Sardinia, 40°33′N, 8°19′E), and South Pacific (Whakatane, New Zealand, 37°31′S, 177°11′E). Infested samples were selected owing to their typical green-to-gray coloration. They were air dried before transport to the laboratory and then rehydrated in a mixture of Provasoli's enriched seawater (PES) medium (19, 20) at pH 8.3 and filtered seawater in a 1:1 (vol/vol) ratio. Using intertidal samples, which naturally suffer desiccation cycles, ensured that microbial activity would resume upon rewetting. Thereafter, samples were kept at 25°C under constant light, provided by white fluorescent lamps at a light intensity of 30 μmol (photons) m⁻² s⁻¹, before analyses. Fragments were analyzed within a week of collection to minimize potential changes in community composition or a significant loss of biological activity.

Confocal laser scanning microscopy.

Small fragments, ranging from 2 to 6 mm² in size, were broken off the samples using sterile pliers and/or a sterilized hammer. Fragments were scrubbed briskly with a small sterile paintbrush to remove any superficial biomass or biofilms, so as to ensure that no epiliths were included in the analyses, rinsed twice in sterile PES medium, and prepared for confocal microscopy analysis as described previously (12). Briefly, fragments were incubated for 1 h in PES medium with 1 μM Calcium Green-5N ([CG5N] Molecular Probes, Eugene, OR) at room temperature under incandescent light at 30 μmol of photons. After incubation, the fragments were fixed to custom-made, slide-mounted chambers, which were then filled with PES containing CG5N at a final concentration of 1 μM. The slides were then placed on the stage of a Leica TCS-SP2 confocal laser scanning microscope and observed under a 40× oil immersion objective. Epifluorescence was used to evaluate the fragments for infestation and viability (presence of photosynthetic pigments in cells) prior to laser scanning. Visible laser lines of Ar/Kr (488 nm) and Kr (546 nm) were used to excite the samples in confocal mode. Confocal laser microscopy allowed us to optically section the areas around the calcite chip surface without physically disturbing any gradients that had formed. Single-plane horizontal fluorescent images were stored in two separate channels: green fluorescence for Ca²⁺ and red autofluorescence for chlorophyll a. Green fluorescence intensity was converted into absolute Ca²⁺ concentrations using a two-point calibration where the signal inside the solid chip corresponded to zero Ca²⁺ and the signal far away from the chip corresponded to the medium in equilibrium with calcite (10 mM Ca²⁺). Curves of concentration versus fluorescence for Calcium Green in PES medium were run separately in a spectrofluorometer (Turner Designs). In PES medium at pH 8.3, the dissociation constant of the dye was around 5 mM, and a linear approximation described the relationship as well as a hyperbolic one in the measured range. The linear relationship was used for extrapolation into the supersaturated region, as it will bias by defect (i.e., it is a conservative concentration estimator [see reference 12]). Dynamic changes in fluorescence with time, as used here, were monitored on a single optical section focused on the surface of the carbonate, in the region around the borehole entrance. Images were analyzed with the Leica Confocal Software (Leica Microsystems Inc., Bannockburn, IL).

Testing for boring activity and its light dependence.

To test for active boring, we imaged [Ca²⁺] in situ at the surface of infested fragments on a single optical section. This area at the borehole entrance has the strongest [Ca²⁺] supersaturation (≫10 mM) in pure-culture experiments using BC008. Fragments were illuminated with the stage illumination (white light, 50 μmol photons m⁻² s⁻¹) for 1 h before initial imaging. To demonstrate light dependence, a characteristic of boring by phototrophs, illumination was turned off for at least 1 h and the Ca²⁺ concentration was imaged again. To assess recovery, the light was turned on and fragments were incubated for another hour before the measurements were repeated.

Effect of calcium transport inhibitors on boring.

To probe the involvement of P-type Ca²⁺-ATPase enzymes in the boring mechanism of field euendoliths, the inhibitors thapsigargin (TG; Alexis Biochemicals, Lausen, Switzerland) and tert-butylhydroquinone (TBHQ; Spectrum Chemicals, Gardena, CA) were used. The initial concentration was 100 μM, which was increased in sequential experiments if negative results were obtained. The experiments using the minimal concentration with a measurable effect are reported. Carbonate fragments were incubated under constant light for 1 h, and single-optical-plane measurements were made at the sample surface before addition of the inhibitors. After initial addition of the inhibitors [Ca²⁺] dynamics were monitored for at least 1 h.

Extraction of euendoliths and microscopic observations.

Euendoliths were extracted from the carbonate samples used in the confocal microscopy experiments described above by an EDTA-based dissolution method based on that of Wade and Garcia-Pichel (25). Briefly, carbonate fragments were washed in sterile PES and placed on a mesh basket on top of the column of a filtration apparatus. There, an iced-cooled, sterile solution of 100 mM EDTA, pH 5, was allowed to drip over the sample at a rate of 0.5 ml min⁻¹, slowly dissolving the carbonate matrix. Cells thus exposed were removed by gently brushing the surface, allowed to flow with the dripping EDTA solution, and collected on a (2-μm-pore-size) polycarbonate filter at the bottom of the column. Filters were washed with sterile distilled water (∼5 ml) while still in the column, collected, and placed in 5 ml of sterile PES medium. After gentle centrifugation (4,000 × g,10 min), filters were removed and cells were resuspended in sterile PES medium. Wet mounts were prepared in glass slides for observation under bright-field optics in a compound microscope. Cell diameter measurements were performed with a calibrated ocular micrometer.

DNA extraction and amplification.

Endolith biomass was extracted by EDTA dissolution as described above (25). Nucleic acids were then extracted from the biomass of the same samples used in physiological experiments and microscopy with the UltraClean Soil DNA kit (MoBio Laboratories, Inc., Solana Beach, CA) according to the manufacturer's recommendations. After extraction, genomic DNA quantity and size were determined by electrophoresis on 1% agarose gels stained with ethidium bromide. Approximately 10 ng of DNA extract was used as the template to generate small-subunit rRNA gene (16S rRNA) amplicons by PCR amplification. 16S rRNA fragments (ca. 700 bp long) were amplified using the primer set CYA106F/CYA781R, which is specific for cyanobacteria (25). The thermal cycle consisted of an initial denaturation at 94°C for 5 min; 35 cycles of 94°C for 1 min, 60°C for 1 min, and 72°C for 1 min; and a final extension at 72°C for 5 min. Quantification of PCR products was performed as explained previously for genomic DNA.

Clone libraries.

To gauge the diversity of euendoliths, clone libraries of the 16S rRNA gene PCR products were constructed separately for each fragment used in the inhibitor experiments. These were generated by ligating 16S rRNA amplicons into a TOPO 2.1 cloning vector and transforming chemically competent E. coli cells by following the manufacturer's instructions (Invitrogen). Ten individual colonies were selected from the library, and each was grown in 3 ml of Luria broth with kanamycin (50 μg/ml) overnight at 37°C in a shaking water bath. A Qiagen Plasmid Prep kit was used to isolate the plasmids according to the manufacturer's instructions (Qiagen Inc.). EcoRI was used to check for the correct insert size. Double-stranded plasmid DNA was sequenced in an Applied Biosystems 3730 sequencer.

Phylogenetic reconstructions.

A total of 11 clones were obtained from Italian samples, 16 from those from Mexico, 10 from those from New Zealand, and 8 from those from Puerto Rico. Sequences were aligned with MEGA 4.0 (22) and checked for noncoding bases, which accounted for 1% or less of the total sequence length. All noncoding ends were discarded. These new sequences were aligned alongside other cyanobacterial 16S rRNA partial sequences retrieved from the National Center for Biotechnology Information database using the basic local alignment search tool. The cyanobacterial sequences most similar to our clones were used to establish the initial alignment, and sequences representative of all taxonomic groups of cyanobacteria (and plant plastids) were added to populate the alignment; a betaproteobacterium sequence was used to root the phylogenetic tree. In all, 93 sequences were used to construct the main phylogenetic tree, which contains all clones from all sites. All positions containing gaps and missing data were eliminated. Two algorithms were used, neighbor joining (NJ) and the maximum parsimony (MP), both with 1,000 bootstrap replicates. Bootstrap values, or the percentage of replicate trees where a group of sequences clustered together, are shown next to the respective nodes.

Nucleotide sequence accession numbers.

Sequences obtained during this work have been submitted to GenBank and assigned accession numbers JN810702 to JN810746.

RESULTS

Morphology of euendoliths.

Morphogenera were identified by comparison with traditional morphotaxa (4, 14). Figure 1 nonexhaustively illustrates their diversity in our samples. In the Italian samples (IT) (Fig. 1, upper left panel), we found pseudofilamentous, nonheterocystous forms, resembling Hyella or Solentia, that ranged from 10 to 15 μm in diameter and displayed a range of bluish-to-greenish colorations. We also found abundant thin, filamentous, nonheterocystous, Plectonema-like forms with trichome diameters of about 2 to 3 μm. In the New Zealand fragments (NZ) (Fig. 1, upper right panel), we detected an unidentified, thin, true-branching, apparently nonheterocystous form 2 to 3 μm in diameter and of uncertain assignment. In this fragment, we also commonly found nonheterocystous, baeocyte-forming Pleurocapsa- or Mysoxarcina-like morphotypes with various vegetative cell diameters (5 to 15 μm). In the Mexican (MX) samples (Fig. 1, lower left panel), nonheterocystous, grayish, Hyella-like pseudofilaments 10 to 15 μm in cell diameter were common, accompanied by 2- to 3-μm-thick, filamentous, nonheterocystous, Plectonema-like types of elongated cells. Other thin, filamentous, nonheterocystous forms with shorter cells 2 to 3 μm in diameter and blue-green in color were seen there as well. Lastly, in the Puerto Rican fragments (PR) (Fig. 1, lower right panel), filamentous, true-branching, heterocystous forms were observed that ranged from 5 to 10 μm in diameter, with the typical morphology of Mastigocoleus, as well as filamentous, nonheterocystous forms with apparent baeocyte formation, ranging from 4 to 10 μm in diameter, whose assignment is more difficult.

Fig 1.

Morphological diversity of extracted euendoliths. Upper left (IT), forms similar to Hyella and Plectonema. Upper right (NZ), Plectonema-like (PL) and Pleurocapsa/Mysoxarcina (PM). Lower left (MX), Hyella (H) and Plectonema (PL) morphotypes. Lower right (PR), unclassified morphotypes, baeocyte-forming pseudofilaments (U), and Mastigocoleus-like filaments (M). IT, Italy; MX, Mexico; NZ, New Zealand; PR, Puerto Rico. Bars, 10 μm.

Boring activity and its light dependence.

Incubations in the light were performed on all fragments to assess if Ca²⁺ supersaturation, a proxy for boring activity by phototrophic euendoliths (12), was reached at the surface of the substrate. All fragments examined displayed this clearly (an example is shown in Fig. 2). Separate killed controls did not. Figure 3 depicts extracellular Ca²⁺ concentrations and their dynamic response to sequential darkening and illumination in the areas around boreholes, as quantified from a series of sequential images such as that shown in Fig. 2. All samples exhibited significant supersaturation, with an average [Ca²⁺] of 20 to 60 mM in the light, and in all cases, darkening resulted in [Ca²⁺] relaxing back to the concentration of the medium in dissolution equilibrium with calcite; for PES medium this is around 10 mM (12). Upon sequential onset of illumination, again in all cases, there was clear recovery of supersaturation levels. Thus, all samples tested contained significant boring activity, most of which can be attributed to phototrophic organisms that require light to power calcite dissolution. This, however, does not necessarily require a biological calcium transport system to be at work, in that similar results would have been obtained, for example, by localized and intense acidification of the medium, which would force the dissolution of calcite and the release of excess metal.

Fig 2.

Surface of a shell fragment (sample from New Zealand, top view) containing a natural assemblage of live phototrophic euendoliths, as imaged by confocal laser scanning microscopy. Green fluorescence intensity denotes extracellular Ca²⁺ concentration as reported by the fluorophore CG5N. Red autofluorescence is from chlorophyll and/or phycobilins. Yellow/orange fluorescence indicates colocalization of both signals. High Ca²⁺ concentrations colocalize with boreholes or near terminal cells of filamentous microborers.

Fig 3.

Dynamics of Ca²⁺ concentrations around the surface boreholes of infested carbonate fragments measured by laser scanning fluorescence microscopic imaging of live samples during shifts in illumination. Dotted lines represent 1 standard deviation of the population mean (solid symbol) of at least 10 measurements. IT, Italy; MX, Mexico; NZ, New Zealand; PR, Puerto Rico.

Effect of Ca²⁺-ATPase inhibitors on boring.

Figure 4 illustrates the effect of the Ca²⁺-ATPase inhibitor TG on the surface [Ca²⁺]. Average Ca²⁺ supersaturation levels on the fragments used for these experiments and before inhibitor addition ranged from 40 to 70 mM. TG was effective in inhibiting boring activity in the NZ and PR fragments, with free Ca²⁺ reaching saturation levels (∼10 mM) soon after the treatment. The IT and MX fragments were apparently not significantly affected, even at high concentrations. Figure 5 illustrates the effects of the specific Ca²⁺-ATPase inhibitor TBHQ. The average Ca²⁺ supersaturation levels in all of the fragments used here ranged from 32 to 75 mM initially. Treatment with TBHQ resulted in effective inhibition of Ca²⁺ supersaturation in the IT, NZ, and PR samples. This effect was dose dependent, with IT responding to 5 mM TBHQ but not to a previous 1 mM treatment (data not shown). The MX samples did not respond to TBHQ (up to 10 mM). Thus, all but the MX samples were sensitive to at least one of the two P-type ATPase inhibitors tested.

Fig 4.

Dynamics of Ca²⁺ concentrations around the surface boreholes of infested carbonate fragments measured by laser scanning fluorescence microscopic imaging of live samples and effect of the addition of the P-type Ca²⁺-ATPase inhibitor TG (1 mM). Dotted lines represent 1 standard deviation of the population mean (solid symbol) of at least 10 measurements. IT, Italy; MX, Mexico; NZ, New Zealand; PR, Puerto Rico.

Fig 5.

Dynamics of Ca²⁺ concentrations around the surface boreholes of infested carbonate fragments measured by laser scanning fluorescence microscopic imaging of live samples and effect of the addition of the P-type Ca²⁺-ATPase inhibitor TBHQ (1 to 10 mM). Dotted lines represent 1 standard deviation of the population mean (solid symbol) of at least 10 measurements. IT, Italy; MX, Mexico; NZ, New Zealand; PR, Puerto Rico.

Clone libraries and phylogenetic reconstruction.

The phylogenetic placement of clone sequences was inferred using the NJ and MP algorithms with 1,000 bootstrap replicates. Both algorithms generated trees with similar topologies; only the NJ tree is shown for the sake of simplicity. Figure 6 illustrates the phylogenetic analysis of 93 16S rRNA partial sequences, including 45 clones, 42 cyanobacteria, 3 plasmid sequences, and 1 member of the betaproteobacteria (Chromobacterium violaceum JCM 1249) as an outgroup. In order to ensure proper correlation, we used the spent samples from the inhibition experiments to obtain 16S rRNA gene clone libraries. A total of 11 sequences were obtained from Italy (Italy sequences), 16 were from Mexico (Baja sequences), 10 were from New Zealand (NZ sequences), and 8 were from Puerto Rico (PR sequences). Eight clades (A, B, B₁, C, D, E, F, and G) were assigned to clone sequences that cluster together. Clade A includes sequences allied to the heterocystous cyanobacteria in taxonomic groups IV (order Nostocales) and V (order Stigonematales) and includes the sequence from model strain Mastigocoleus sp. strain BC008. Clade B includes baeocyte-forming genera (Mysoxarcina/Pleurocapsa/Staniera) in group II (order Pleurocapsales). Clade B₁ contains sequences that fall within clade B but are not well resolved and do not have any close neighboring sequence. Clade C contains group II sequences and includes Chroocococcidiopsis and Solentia sp. strain HBC10, a demonstrably boring cyanobacterium (8). Clade D includes a cluster that falls within group II but does not have a closely allied sequence in databases. Clades E and F contain sequences exclusive to the New Zealand sample which fall within group III (order Oscillatoriales; nonheterocystous and filamentous). Clade E contains Leptolyngbya types (also nonheterocystous and filamentous), with its closest neighbor being Leptolyngbya sp. strain ITAC101 isolated from a marine sponge; clade F does not have a closely matching sequence in the databases. Clade G includes Leptolyngbya-type sequences (group III), with the closest neighbor being Leptolyngbya sp. strain HBC1, an isolate from marine carbonate microbialites (8). Apart from the assigned clades, there are three clone sequences, all likely allied to Leptolyngbya that stand alone: PR 10e, Baja 5h, and Baja 6a. From these, only Baja 5h has a close neighboring sequence pair (Leptolyngbya sp. strain PCC 7373 and Phormidium sp. strain SAG 80.79).

Fig 6.

NJ phylogenetic tree based on 93 16S rRNA gene partial sequences (559 bp). The sequence of C. violaceum JCM 1249 was used as an outgroup. Letters denote clades of euendolithic phylotypes. Colors indicate the geographic regions of origin (red, Italy; green, Mexico; magenta, New Zealand; blue, Puerto Rico). Denominations of entries correspond to those given in the database and are not necessarily taxonomically correct. Bootstrap values for 1,000 trees are indicated at the nodes. The scale bar represents 1% estimated sequence divergence.

DISCUSSION

Microborer diversity.

Microscopic observations of extracted euendoliths revealed a broad diversity of morphogenera in our sample collection. While we acknowledge that present taxonomic divisions of the cyanobacteria offer just a convenient, rather than a phylogenetically correct, view of diversity, it is still noteworthy that the euendoliths detected here encompassed 3 out of the 5 major cyanobacterial taxonomic groups (group II, order Pleurocapsales; group III, order Oscillatoriales; group V, order Stigonematales), excluding at this time groups I and IV, Chroococcales and Nostocales, respectively. Most common were pseudofilamentous, nonheterocystous, baeocyte-forming cyanobacteria (group II), which could be seen in all four of the geographical regions surveyed. This result is not surprising, as it is from this group II of the cyanobacteria that many morphogenera with microboring species, such as Cyanosaccus (17), Hyella (3, 16), Hormathonema (7, 14), and Solentia (17, 18) have been described. The importance of this group is particularly evident in the sard samples, where all of the forms we found can be assigned to it, at least morphologically. In the fragments from New Zealand and Mexico, forms assignable to groups II and III were found. In Mexico, we found Hyella/Solentia-like cyanobacteria among others that closely resembled members of the genus Plectonema. In the New Zealand samples, we found Pleurocapsa/Mysoxarcina-like forms among Plectonema-like types. Plectonema is another genus that contains boring species (e.g., Plectonema terebrans; 4, 15). In the fragments from Puerto Rico, extracted filaments displayed a variety of morphologies, with many (pseudo)filamentous, nonheterocystous types capable of apparent baeocyte formation (likely group II), but we were unable to assign them accurately to an existing morphogenus. We also detected filamentous, true-branching forms that displayed lateral heterocysts (group V, order Stigonematales) closely resembling the morphological characteristics of Mastigocoleus, another genus with a single, microboring species, M. testarum.

This broad morphological diversity is partly consistent with the variety of clades detected according to phylogenetic reconstructions of 16S rRNA clone libraries. Four out of eight clades found (B, B₁, C, and D) were assigned to clusters of sequences that fall within baeocyte-forming (group II) types. The relative abundance of sequences within this group, compared to all other phylotypes, indicates that the majority of euendoliths found are pseudofilamentous, nonheterocystous, and baeocyte forming, which correlates with the common microscopic observations of Hyella- and Solentia-like forms and other baeocyte formers. Clade A includes one sequence found only in Puerto Rico, and the sequence of the model boring organism M. testarum BC008, which was originally isolated from carbonates in this geographic region (5). This finding is not surprising, as it correlates fully with microscopic observations. The only conspicuous disagreement is in the relative diversity and abundance of Leptolyngbya-like, thin-filamentous cyanobacteria. Traditional studies have reported only a single species of microborer among this group: Plectonema (Leptolyngbya) terebrans (4, 14, 15). Yet we found at least three distinct clades in our survey (E, F, and G) and a few singletons, which likely represent several different generic entities. Obviously traditional morphological surveys have severely underestimated the diversity of thin-filamentous microborers. This is perhaps not surprising, since other studies have shown that much of the “hidden” genetic diversity in cyanobacteria resides in the morphologically simplest forms, either filamentous or unicellular (10). Our phylogenetic data suggest that many species, perhaps even genera within group III (Oscillatoriales), are capable of boring and that this diversity remains largely unexplored. No cultivated isolates of this type exist currently in public culture collections. Some of the stand-alone clone sequences from Mexico and Puerto Rico might correlate with their respective yet-to-be-identified morphotypes. Thus, a renewed and sustained effort to find cultivated isolates for some of these clades and unidentified forms is necessary to fully describe the diversity of euendoliths in a comprehensive and useful manner.

Universality of the boring mechanism.

Having established that all samples were active with respect to boring activity and that phototrophs accounted for most of the boring activity in these substrates, the molecular surveys discussed above ensured that the diversity of microborers was large in all samples and covered many of the groups of cyanobacterial microborers known from the literature. Thus, the fact that in most instances at least one of the P-type Ca²⁺-ATPase inhibitors could abolish boring activity indicated that these enzymes are central to the process of boring in most cyanobacterial euendoliths, as they are in strain BC008. In this sense, the mechanism seems to be quite widespread. Not unexpectedly, however, the degree of sensitivity differed among samples. We knew from laboratory experiments that inhibitors such as TG can work in some organisms and not in others (13), and this is why our experimental design called for the use of two independent inhibitors. In our interpretation, sensitivity to at least one of the two implies the involvement of the target enzymes. Only in one case, the samples from the Sea of Cortez (MX), there was no response to either TG or TBHQ, even at high concentrations. This prevents us from calling the response universal. The insensitivity may speak for the presence of a different boring mechanism in the euendoliths of this sample. If so, then we would need to postulate that a group of cyanobacteria that is highly represented in the Mexican clone libraries, but not in others, is probably responsible for the alternative behavior. Clade B₁ seems to be the only group fitting this criterion. Alternatively, clade B₁ may possess a form of calcium transporters that are particularly insensitive to the inhibitors used. Unfortunately, no cultivated representatives of this clade are yet available to test the hypothesis of functional differences.

The empirical analysis of the boring mechanism carried out here with geographically distinct, complex euendolithic microbial assemblages reveals similar physiological response to core experiments performed with M. testarum BC008, suggesting that the boring mechanism, as it is understood in the latter, is probably widespread, if perhaps not universal. Our work also revealed that the genetic diversity of cyanobacterial euendoliths, particularly the nonheterocystous filamentous form, has been underestimated in the past and provides a phylogenetic basis for the study of cyanobacterial euendolith biodiversity.