Extreme cellular adaptations and cell differentiation required by a cyanobacterium for carbonate excavation

Significance

Microorganisms can play a significant role as agents of erosion. Euendolithic cyanobacteria, photosynthetic microorganisms that bore their way into carbonate substrates, are among the most important bioerosive agents known. The mechanism by which they dissolve and excavate carbonate rocks has remained a mystery for decades, but is now known to involve calcium transport from cell to cell. Herein we show that the mechanism involves two adaptations unprecedented in prokaryotic biology: the polar placement of calcium and proton transporting enzymes in a manner that is coordinated between cells and the differentiation of a novel type of cell that is able to accumulate calcium at levels some 500-fold those found in normal cells.

Abstract

Some cyanobacteria, known as euendoliths, excavate and grow into calcium carbonates, with their activity leading to significant marine and terrestrial carbonate erosion and to deleterious effects on coral reef and bivalve ecology. Despite their environmental relevance, the mechanisms by which they can bore have remained elusive and paradoxical, in that, as oxygenic phototrophs, cyanobacteria tend to alkalinize their surroundings, which will encourage carbonate precipitation, not dissolution. Therefore, cyanobacteria must rely on unique adaptations to bore. Studies with the filamentous euendolith, Mastigocoleus testarum, indicated that excavation requires both cellular energy and transcellular calcium transport, mediated by P-type ATPases, but the cellular basis for this phenomenon remains obscure. We present evidence that excavation in M. testarum involves two unique cellular adaptations. Long-range calcium transport is based on active pumping at multiple cells along boring filaments, orchestrated by the preferential localization of calcium ATPases at one cell pole, in a ring pattern, facing the cross-walls, and by repeating this placement and polarity, a pattern that breaks at branching and apical cells. In addition, M. testarum differentiates specialized cells we call calcicytes, that which accumulate calcium at concentrations more than 500-fold those found in other cyanobacteria, concomitantly and drastically lowering photosynthetic pigments and enduring severe cytoplasmatic alkalinization. Calcicytes occur commonly, but not exclusively, in apical parts of the filaments distal to the excavation front. We suggest that calcicytes allow for fast calcium flow at low, nontoxic concentrations through undifferentiated cells by providing buffering storage for excess calcium before final excretion to the outside medium.

Cyanobacteria that actively bore into carbonaceous substrates (also known as euendoliths) are a functional class of microbes that contribute to the erosion of marine (1, 2) and terrestrial (3) carbonates, both mineral and biogenic. These organisms can affect coral reef ecosystems and constitute a commercially significant pest to bivalve farms (2, 4). Carbonate excavation rates of euendolith communities may increase with elevated oceanic pCO₂, as is the case with boring sponges (5, 6). Despite their global environmental effect, and their potential sensitivity to global climate change, the physiological and cellular mechanisms behind photoautotroph-mediated carbonate dissolution have remained largely elusive (7). Cyanobacteria, similar to other autotrophs, alkalinize their local microenvironment by virtue of their carbon fixation activity, typically facilitating carbonate precipitation (8), not dissolution. In the case of marine euendoliths, seawater is supersaturated with respect to calcite and aragonite, making carbonate dissolution thermodynamically unfavorable. Both types of constraints render the process of boring quite paradoxical (7, 9).

In our laboratory, we have recently developed a conceptual model of the physiology of cyanobacterial carbonate excavation based on experiments carried out with the model filamentous euendolith, Mastigocoleus testarum strain BC008 (10). Similar to many filamentous cyanobacteria, this strain is a true multicellular microbe: it displays complex morphological plasticity, which includes cell differentiation of lateral heterocysts dedicated to nitrogen fixation, motile multicelled hormogonia for dispersal, true branching of the main thallus, and the ability to alter filament morphology during boring into solid substrates (11). In the current model (10), calcium ions are preferentially imported by passive diffusion into apical cells that are proximal to the boring front. This lowers the Ca²⁺ concentration in the interstitial microenvironment between this apical cell’s surface and the solid mineral, below the solubility product of the corresponding mineral, thus shifting equilibrium there toward dissolution. This allows the organism to selectively dissolve the substrate directly adjacent to the apical tips of boring filaments and to grow into that space. Intracellular calcium (Ca²⁺_i) then moves cell to cell from the boring front toward the mineral surface, where it is actively pumped out into the liquid medium by distal apical cells. Extruded excess Ca²⁺ can either diffuse away or, in some cases, reprecipitate into a micrite rind (12, 13). Boring continues into the solid substrate until the organism reaches its limit of light intensity for growth.

The model called for the agency of P-type calcium ATPases in the transport process, but did not address whether the pumps are predominantly located on the distal cells, where extrusion to the medium occurs, or whether they are also located between cells, thus participating in the cell-to-cell transport of Ca²⁺. The model also assumed that protons constitute the counter ions exchanged for calcium to maintain charge. The concentrations of Ca²⁺_i attained during transport, which are of paramount importance for cell physiology, given the ion’s toxicity in most cellular systems, including cyanobacteria (14, 15), remain unknown, but there is a potential that a borer’s cells might experience prolonged, very high Ca²⁺_i concentrations (10). The experimental evidence in support of the current model was obtained using selective inhibitors and an extracellular calcium fluorophore to study Ca²⁺ dynamics during boring. Although this method was appropriate to assess bulk Ca²⁺ transport, it did not address intracellular processes directly. In this contribution, we intended to assess Ca²⁺ transport and related processes in M. testarum, by extending experimentation to the intracellular domain during both boring and planktonic growth.

Results and Discussion

Vectorial Transfilamentous Calcium Transport.

If transport were driven exclusively by distal apical pumps, then severing a filament midway should not result in leakage of calcium at the breakpoint, given that the extracellular concentration in seawater medium (10 mM) is much higher than physiologically plausible (15). However, if pumps existed along the entire filament, they would remain active, and localized regions of calcium supersaturation in the medium would be observed adjacent to filament breakpoints, just as it is seen close to the apical cells that extrude Ca²⁺ to the outside medium (10). Mechanical severing of boring filaments at various filament locations after making them accessible by cleaving the mineral substrate along the direction of excavation did indeed result in substantial and localized increases in Ca²⁺ in regions directly adjacent to damaged filaments (Fig. 1B, white arrowheads), as measured with extracellular Ca²⁺ dyes. To continually “bleed” Ca²⁺, as observed, a back pressure of calcium must have existed. This pressure can only be supplied by continuous active pumping of an adjacent cell into the severed cell (or the space around the severed cell). Thus, Ca²⁺ pumping occurs along the length of each boring filament, using many active pumps located in more than one cell within each filament, clearly invalidating the notion that transcellular pumping is driven by a “negative pressure” established at the distal apical cells.

Fig. 1.

Calcium pumping occurs along boring trichomes. Infested calcite chips were cleaved at a plane orthogonal to the surface to expose boring filaments lengthwise along the newly exposed face, which was then scraped with a needle, incubated with a calcium-sensitive dye, and imaged. (A) Transmitted light image showing boring filaments within the chip on the cleavage plane. The original surface was to the upper right, with boring proceeding toward the bottom left. Dashed lines denote scouring lines. (B) Confocal imaging of the same area, where photosynthetic pigment autofluorescence is red and calcium reporter dye is green. Arrowheads point to localized supersaturation in extracellular calcium. When cells were bleach-killed, there was no significant calcium release. (Scale bars, 10 µm.)

P-Type Calcium ATPase Localization.

The inhibition of calcium extrusion in BC008 by P-type calcium ATPase-binding antagonists, such as thapsigargin (10, 16), had revealed their agency in the process of calcium extrusion during boring. In an attempt to localize the exact sites of activity of these enzymes, we used fluorescently labeled thapsigargin (BDTH), which has been shown to bind specifically to eukaryotic P-type calcium ATPases (17, 18). Controls using only the fluorophore [boron-dipyrromethene-FL (BODIPY-Fl), with no thapsigargin] showed some nonspecific binding to the extracellular sheath, but no particular binding patterns, and no membrane colocalization (Fig. S1B). A control with thapsigargin alone showed no fluorescence in the appropriate channel. M. testarum strain BC008 stained with BDTH displayed homogenous membrane staining, but significantly higher signal was observed close to the cell septal region (Fig. 2A). Three-dimensional reconstructions of the optical sections show that the regions of increased staining manifest as an annulus adjacent to the septal area (Fig. 2B). Importantly, only one of the two septal sides in a given cell showed preferential staining, and the polarity of the annuli tended to be repetitive along neighboring cells. Separate experiments using membrane-specific dies showed that our current technique allows us to resolve both sides of the septa (Fig. S1A). This clearly suggests a sequential, vectorial transport in a single, coherent direction along the filament. This coordinated polar pattern seems to break down at filament branches, or at certain regions of adjacent cells containing very low photopigment autofluorescence (PSAF), in which signals are intense all around the membrane (Fig. 2A, asterisk). In some of these special regions, one can indeed see two zones of intense BDTH staining at opposite sides of the same septum. Boring filaments that had been disinterred from the calcite matrix showed indistinguishable patterns of binding to those seen in mineral-free filaments. However, the intensity of BDTH signal in boring biomass was significantly higher (six times; P < 0.03, t test), indicating an increased presence of the transporters in membranes of boring cells (Fig. S2). This corroborates previous findings at the gene expression level (10). As P-type ATPases extrude calcium, they antiport protons (19) and are likely to alter pH locally if found in sufficient concentration. We used a ratiometric pH-sensitive fluorophore to monitor intracellular variations in pH that may denote their activity and preferred location. In agreement with the localization observed by direct staining, we could observe clear periplasmic alkalization adjacent to septal annular regions (Fig. 2C). Together, these observations support the central role of P-type calcium ATPases in the excavation process, indicating the presence of a vectorial arrangement of these pumps in vegetative cells that drives transcellular transport.

Fig. S1.

General organization of the cellular envelope of M. testarum BC008. (A) Localization of the continuous outer membrane (yellow/green channel, stained with FM4-64FX). This image represents a single optical equatorial slice in a mineral-free grown filament. Blue fluorescence is a result of DAPI staining of DNA. The cell septae (as well as the sheath) are autofluorescent, allowing us to easily discern both sides of the cell cross-wall, showing that our technique can resolve membrane staining at either side of the septum. (Scale bar, 1 μm.) (B) The BODIPY-FL moiety alone does not show localization to the membrane, but diffusely to the sheath region, as indicated by the large distance from the intracellular PSAF signal (white arrowhead). (Scale bar, 10 μm.) (C) Transmission electron microscopy image of a cell septum from a typical mineral-free BC008 filament showing a septal thickness of 35–45 nm. (Scale bar, 50 nm.)

Fig. 2.

Cellular localization of P-type calcium ATPases by thapsigargin targeting. (A) Most cells present strong preferential labeling (green) at one side of the septum (white arrowhead), with the polarity maintained from cell to cell. Special cells (asterisks) show a generalized, strong labeling appearing on both sides of adjacent septa. (Scale bars, 10 µm.) (B) Three-dimensional reconstructions reveal that the preferential membrane label forms an annulus facing the periphery of the septal region. Controls using unconjugated fluorophore (Fig. S1B) or thapsigargin alone showed no membrane binding. (C) Ratiometric pH image showing increased pH at septal periplasmatic regions facing the ATPase annuli. Absolute calibrations were carried out using Nostoc cultures.

Fig. S2.

Differential concentration of septal P-type calcium ATPases as a function of growth mode. This was gauged as the ratio of bound BDTH signal to PSAF signal (a proxy for Chl a) obtained from background-corrected summed Z-projections of six independently obtained image stacks (three boring and three mineral-free) under the same imaging conditions (boring biomass was disinterred from the mineral before staining). The ratio was then corrected by the cellular content of chlorophyll a typical of each growth mode, determined separately, as a ratio to protein content. The concentration of septal P-type calcium ATPases was significantly higher in boring cells (*P < 0.03). RU, relative units.

Intracellular Calcium Dynamics.

Cyanobacteria show a very tight regulation of Ca²⁺_i levels, typically in the low nanomolar range (15, 20, 21), but Ca²⁺ pumping in euendoliths could potentially create transient or even continued intracellular hyperaccumulation (10). We used several approaches to probe Ca²⁺_i in strain BC008. In our initial studies, we used the extracellular indicator calcium green 5-N (CG5N), noticing that after prolonged incubation, it could penetrate filaments of BC008. Using CG5N with transversally cracked calcite chips containing intact filaments (Fig. 3E), Ca²⁺_i accumulation was clearly observed in cells corresponding to apical distal tips as they extended into the liquid medium. However, some groups of cells inside the chip did also show such accumulation. Interestingly, these high-calcium cells, which we call calcicytes, presented visibly less photosynthetic pigment (judged by their low PSAF signal) than cells with low Ca²⁺_i. This diagnostic inverse PSAF-to-calcium signal of calcicytes was not exclusive of apical distal areas, but was also found in specific stretches along filaments, in individual cells, or more often than not, in a group of adjacent cells. Similar results were obtained using the intracellular Ca²⁺ reporter dye Fluo-5F AM both in cultures boring inside calcite chips (Fig. S3B) and those grown planktonically (Fig. S4). To determine whether this marked cellular heterogeneity in Ca²⁺_i was an unusual trait specific to BC008 or was common to cyanobacteria at large, we extended our analyses to two common but not endolithic laboratory strains of filamentous cyanobacteria: Nostoc punctiforme 29133 and Microcoleus vaginatus strain PCC 9802. These two strains showed very low Ca²⁺_i signal compared with that obtained in strain BC008, and very small cell-to-cell variations in Ca²⁺_i (Fig. 3 B and C). M. testarum BC008 clearly shows a high degree of cellular heterogeneity of both Ca²⁺_i content and photosynthetic pigment complement. A quantitative analysis of the average volumetric calcium and PSAF intensities in cells of BC008 shows the presence of two distinct cell populations. One encompasses cells with PSAF and calcium similar to Microcoleus or Nostoc (Fig. 3D). The second population, which typically includes cells close to the tips of filaments (Fig. 3D, colored squares), has high Ca²⁺_i signals; in cells from this population, there is a linear inverse relationship between Ca²⁺_i and photopigment, indicative of progressive spatial separation between photosynthesis and Ca²⁺_i hyperaccumulation. It can be logically postulated that high Ca²⁺_i levels may interfere with photosynthesis, and that there is a causal relationship in this correlation. We note that hyperaccumulating cells are clearly neither dead nor moribund: intracellular dyes are only fluorescent in the presence of active cellular esterases (22, 23), and all cells contained intact DNA (as determined by DAPI staining, not shown), which speaks for the integrity of the cell membrane. Thus, the most parsimonious explanation for these findings is that cells in distal apical (and some intercalary) regions may be differentiating into specialized calcium Ca²⁺_i holding cells, or calcicytes.

Fig. 3.

Cellular heterogeneity in calcium concentration. Confocal projection images of photosynthetic pigment autofluorescence (red) and Ca²⁺_i reporter fluorescence (Fluo-5F-AM, green) in mineral-free cultures of BC008 (A), N. punctiforme ATCC 21933 (B), and M. vaginatus PCC 9802 (C). (D) Quantitation of Ca²⁺_i in cells of BC008 (n = 138), Nostoc (n = 10), and Microcoleus (n = 113) plotted against their photosynthetic pigment content. Colored squares denote apical cell groups (BC008 only). (E) Confocal projection, as in A, of a calcite chip infested by BC008, split orthogonal to the surface to reveal a cross-section of the boring bed. The dashed line represents the interface between the medium (above) and the solid (below). (Scale bars, 10 µm.)

Fig. S3.

Intracellular calcium control images. Confocal images (summed stacks) of Fluo-5F (green) and PSAF (red) signal in a typical mineral-free culture with no dye but with dye carrier DMSO as control (A) and a freshly cleaved infested chip showing a cross-section of the newly exposed boring bed, which was incubated in 50 μM Fluo-5F AM (B). Long stretches of calcicytes and newly exposed apical boring tips (white arrowheads) can be seen intermixed with nonaccumulating photosynthetic filaments, as well as an external protruding filament showing high calcium signal (white asterisk). (Scale bar, 10 μm.)

Fig. S4.

Mineral-free Microcoleus testarum BC008 incubated in Fluo-5F illustrating reproducibility of observed Ca²⁺_i patterns. The graph shows a quantitation of individual cellular volumetric calcium concentration vs. photosystem autofluorescence. The biomass shows similarly diverging cellular populations as those in Fig. 3. (Scale bar, 10 μm.)

The levels of Ca²⁺_i observed in calcicytes are unprecedented. Average Ca²⁺_i signals in M. testarum (including cells with low and high calcium content) were sixfold higher than those in Nostoc and threefold higher than those in Microcoleus under identical imaging and staining conditions. Maximal Ca²⁺_i could reach 22- and sixfold higher, respectively (Fig. 3D). Although we could not quantitate Ca²⁺_i in BC008 directly, the resting Ca²⁺_i concentration in Anabaena is around 200 nM (13). Assuming this level to be typical for filamentous cyanobacteria, some calcicytes of nonboring BC008 biomass accumulated 4–5 µM Ca²⁺_i and boring calcicytes upward of 100 μM, which is ∼500-fold the resting levels seen in canonical Nostoc cells.

Intracellular pH Heterogeneity.

Assessing intracellular pH (pH_i) in BC008 was imperative to establish a direct connection among transport of calcium in normal cells, its accumulation in calcicytes, and the mechanism of known calcium ATPases (19). Although homeostatic regulation of pH_i in cyanobacteria is not completely understood, they do not tolerate large drops in external pH (24, 25), and those studied keep circumneutral pH_i, even under alkaline external conditions (26, 27). We used a ratiometric pH_i indicator [2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester; BCECF-AM] to interrogate relative pH_i variability in BC008, and could indeed document clear alkalinization within many of the cells containing low PSAF, including distal apical cell groups (Fig. 4), indicating that they, in fact, correspond to calcicytes, consistent with the predicted alkalinizing effect of a calcium/proton antiport. Although our attempts to directly calibrate the pH_i data in BC008 were unsuccessful, comparisons with absolute calibration curves carried out in N. punctiforme were used as a reference. In nonboring cultures, we could detect alkalinization (ΔpH = 1) in cells fitting the low PSAF diagnostic for calcicytes, including apical cell groups (Fig. 4 A and B, arrows). Surprisingly, in these cultures, vegetative cell pH was kept slightly acidic. In boring cultures, cellular variations in pH were much larger (ΔpH = 3), with calcicytes reaching pH above 9.5 (upper limit for reliable measurements). In excavating filaments, cells of acidic pH were closer to the boring front (Fig. 4 C and D), in agreement with an overall directional counter transport of protons. The magnitude of such pH_i variations is consistent with those seen in Ca²⁺_i. Calcicytes thus sustain severe alkalization of the cytoplasm. Incidentally, given that the fluorescent yield of most calcium fluorophores is strongly quenched by increasing pH (10), a second corollary of a calcicyte’s high alkalinity, the levels of Ca²⁺_i we inferred for them are likely underestimated.

Fig. 4.

Ratiometric confocal assessment of intracellular pH. (A and B) Mineral-free cultures, in which arrowheads point to apical filament regions (calcicytes). (C and D) Boring, in-chip filaments (asterisks mark calcicytes). (Scale bars, 10 μm.) pH_i calibrations were based on Nostoc cells.

Conclusion

Our results shed light on the cellular basis of microbial carbonate excavation. Together, they point to a sophisticated mechanism that involves two types of unique cellular adaptations: marked, coordinated, cell polarity in calcium transport within normal cells, and the differentiation of a specialized cell type, the calcicyte, which accumulates large concentrations of Ca²⁺_i and sustains significant alkalization of the cytoplasm, likely to the detriment of their photosynthetic capacity. We cannot ascertain, at this time, whether high Ca²⁺_i or pH_i simply impedes photosynthesis, leading to degradation of photosynthetic pigmentation, or whether there is a morphogenetic regulatory mechanism at play. We postulate that this euendolithic cyanobacterium maintains Ca²⁺_i levels low and pH_i circumneutral (maybe slightly acidic under boring conditions) in most cells, even in the face of high transport fluxes, by concentrating excesses in a few calcicytes that serve as “capacitors,” or depots for the long-range transport of Ca²⁺ between mineral and outside medium. Calcicytes are located at crucial points in the flow or, typically, at the distal apical areas, in preparation for final excretion. If a calcicyte’s phototrophic capacity is fully impeded, it must rely on resources imported from vegetative cells for energy and fixed carbon. The cellular basis of calcium transport seems to rely on a biased distribution in the localization of P-type calcium ATPases, across the filaments, which preferentially pump into the periplasm at one side of the cell. For this to work, one has to postulate the presence of calcium channels for passive diffusion into the next cell; we do not have evidence of this yet. Such a distribution, with ATPases at one side of the cell and channels at the other, if repetitive, could easily account for vectorial transport while maintaining low intracellular concentrations. Direct cell-to-cell transport is less likely, as it would have to cross large spaces that span two cell membranes and one peptidoglycan-rich periplasm (the septum is between 30 and 40 nm thick, and the outer membrane does not extend into the septal region; Fig. S1 A and C). A direct transport would fail to explain the localized pH excursions seen in Fig. 2C.

Some aspects beyond the presence of calcium channels remain to be studied in any detail, particularly how the initial mineral calcium enters the cell past the cell envelope (through the sheath, across two membranes, and the periplasm) at the apical cells in the boring front, and how they are fully extruded at the apical cells. As far as we know, such transcellular, vectorial calcium transport is unique to euendolithic cyanobacteria and represents a novel physiological adaptation in these organisms. It is clear that pH_i dynamics are complex in this organism and directly linked to calcium traffic, presumably in the opposite direction of calcium, toward the boring front. This would facilitate crystal dissolution by allowing the continued conversion of carbonate anions to bicarbonate and CO₂, which could be used as a carbon source in autotrophy within the crystal matrix. Potentially, Na⁺ may directly or indirectly adjuvate the proton counter transport by neutralizing any charge imbalance. Each filament would then be made up of repeating units of polar cells, with one side allowing calcium entry (boring side) and the other mediating calcium release (and subsequent proton entry). This model is particularly appealing, as it may also explain the ability of single cells (3) to bore, given that such a single cell could not penetrate the mineral without some type of calcium transport polarity.

That the evidence for traits necessary for boring, such as the polar localization of calcium transporters (through thapsigargin staining), calcicyte differentiation, and marked Ca²⁺_i/pH_i heterogeneity among cells, was observed not only in boring cultures but also in mineral-free filaments indicates that the boring capacity is constitutive, always on, in M. testarum. The differences found between the two growth modes were only in intensity. This likely reflects the fact that these organisms are found exclusively in the boring form in natural settings, so there is no fitness value in a hardwired induction. It may be that the only mineral-free form of this organism in nature is that of hormogonia during dispersal. We note here that, in agreement with this view, M. testarum strain BC008 cultured without carbonate easily loses its ability to bore. In any event, the sophistication of true multicellular differentiation and division of labor in a single prokaryotic organism is remarkable. The differentiation of calcicytes comes in addition to heterocyst formation, true branching, coordinated changes in trichome width, and formation of motile hormogonia for dispersal, all present in BC008 (10). Because we have a rather significant fossil record of euendolithic cyanobacteria dating back 1.5 Ga (28), one can infer that such true multicellular (i.e., beyond morphological) (29) adaptations in cyanobacteria are quite ancient as well.

Materials and Methods

Cultures and Growth Conditions.

Mineral-free cultures of M. testarum strain BC008 used for experimentation were grown in standard, vented-cap, tissue culture flasks containing 15 mL PES-30 medium, which consists of 30 g Instant Ocean, 10 mM Hepes, and 10 mL a PES nutrient solution stock (30) per liter of distilled water adjusted to a pH of 8.1. Cultures were incubated on a slowly rocking platform, using a 16 h light/ 8 h dark diel illumination cycle, at a light intensity of 22 μmol photon m⁻²⋅s⁻¹. Boring cultures were obtained by adding ethanol-sterilized calcite chips (blocky calcite, Ward’s Scientific) of small size (< 5 mm³) to mineral-free cultures until the chips were colonized. Cultures of N. punctiforme ATCC 21933 and M. vaginatus strain PCC 9802, used for comparisons, were obtained from stocks grown in liquid BG11° medium and Jaworski’s medium (31), respectively.

Confocal Microscopy and Image Analyses.

All laser scanning confocal fluorescence microscopy was performed on either a Leica SP2 or a Leica SP5 confocal microscope. Images were obtained at a minimum of 1,024 × 1,024 resolution (unless specifically noted), with a minimum line average of 2 and a scan rate maximum of 400 MHz.

Specimen preparation.

All mineral-free samples were imaged live, on glass slides under coverslips, with sealed seams, using a 20× or 63× oil immersion objective. Samples of mineral-boring cyanobacteria were observed in situ while placed in 65 × 25-mm dishes containing 10 mL PES-30 medium and using a 20× or 40× dipping objective.

Fluorescent microscopy and indicators.

Fluorescence emission was measured using different excitation wavelengths as needed. Excitation was performed sequentially to prevent autofluorescence bleed-through into different fluorophore channels. The “red channel” of emission in all images represents the natural photosynthetic PSAF signal and was obtained using an excitation of 561 nm and collected between 660 and 700 nm. The following commercial (Molecular Probes) indicators were used: CG5N, Fluo-5F AM, FM4-64Fx, and BCECF-AM. Hydrophilic indicator stock solutions were prepared at 1-mM concentrations in 10 mM Tris⋅HCl buffer at pH 7.5, and all hydrophobic stocks were prepared (1 mM) in anhydrous DMSO. BODIPY FL-Thapsigargin (BDTH) and BODIPY-FL were also purchased from Molecular Probes and resuspended in anhydrous DMSO at stock concentrations of 2 mM. Acetoxy-methyl ester dyes are cleaved by active esterases in viable cells, which releases the respective fluorophore (32). See Table S1 for use parameters.

Table S1.

Fluorophore use parameters

Indicator	Detection of	Excitation (nm)	Emission (nm)	Final (μm)
CG5N	Extracellular free Ca2+	488	530–540	1
Fluo-5F-AM	Intracellular free Ca2+	496	510–530	25*, 50†
BCECF-AM	Intracellular pH	458/488	525–535	25
BD-TH	P-type Calcium ATPases	496	505–515	20
BODIPY-FL	N/A	496	505–515	20
FM4-64Fx	Lipid Membranes	561	720–740	20

^*For mineral-free cultures.

^†For boring cultures.

Image Analysis.

All image analyses were performed using the MBF 64-bit distribution of the image analysis software ImageJ (33). For each emission channel, we acquired individual Z stacks of images. Each channel stack was background normalized by setting the lower pixel intensity value to that which was observed in stacks obtained from appropriate unstained controls, using identical imaging conditions. Linear contrast optimization was performed on a case-by-case basis and applied equally to all stacks analyzed simultaneously. As needed, normalized images in a stack were then summed, merged, and projected into a single image. To assess relative concentration of analytes, regions of interest were manually circumscribed around cells of interest in a summed stack, and the average value of the area was divided by the cell’s thickness, yielding a value with units relative fluorescence unit per cubic micron (RFU/μm³). The spectral characteristics of the lipid membrane binding dye used (FM4-64FX) overlap with the endogenous pigmentation (PSAF) emission of BC008; thus, to visualize the outer lipid membranes, the PSAF channel was used to create a binary mask. This mask was then used to remove PSAF signal from the FM4-64FX channel, leaving behind only pixels that correspond to emission from dye bound to the outer cellular envelope. To compare the total amount of BODIPY–thapsigargin binding between boring and mineral-free biomass, we processed the image stacks as previously described and measured the total pixel average (measurement of all image pixels) BDTH signal intensity for the normalized, summed, projected stacks by the average total PSAF signal intensity to account for differences in biomass between each image series to obtain a ratio of BDTH signal to PSAF.

Calcium Imaging.

CG5N is canonically used as an extracellular dye in eukaryotic cell biology (34); however, we observed that extended incubation time (≥1 h) of M. testarum strain BC008 in CG5N resulted in intracellular CG5N signal. Actively excavated mineral chip fragments were placed in 10 mL PES-30 with 1 µM CG5N and incubated for at least 1 h under standard culture conditions before imaging. In an effort to validate the observed intracellular CG5N signal, we choose to use a second calcium fluorophore with a lower K_d value than CG5N, Fluo5F-AM (2.3 μM, as opposed to 5 mM), which is typically used as an intracellular calcium indicator. This dye also enabled interspecies comparisons, as CG5N did not seem to penetrate other cyanobacteria. Mineral-free biomass or single chips containing actively boring biomass were placed in 1.5-mL microcentrifuge tubes and rinsed with fresh PES-30 two times. Cultures were then resuspended in 300 μL fresh medium containing 25 μM Fluo5F-AM (50 μM for boring cultures). The cells were incubated for at least 1 h at room temperature and then washed twice in fresh media. Laser scanning confocal fluorescence microscopy and image analysis were performed as previously stated.

P-Type Calcium ATPase Localization.

Thapsigargin is a pharmaceutical found to inhibit P-type calcium ATPases (18) that is also a potent antagonist to the enzymes responsible for calcium transport and excretion during carbonate boring in M. testarum strain BC008, as well as in other euendolithic cyanobacteria in nature (10, 16). In an effort to determine the cellular localization of these P-type calcium ATPases responsible for Ca²⁺ transport during excavation, we used a fluorescently labeled analog of thapsigargin (BODPIY-FL, BDTH) that retains its binding specificity to the transporters in vivo in animal systems (35). For this, mineral-free biomass was placed in a 1.5-mL microcentrifuge tube, rinsed with fresh PES-30 twice, and incubated in 300 μL fresh medium containing 20 μM BDTH for 3 h at room temperature in the dark. Samples were then spun and washed twice with fresh media before imaging, as described earlier. Controls were run using the fluorescent carrier alone (DMSO), thapsigargin alone (no BODIPY), and BODIPY-FL alone (no thapsigargin). Boring biomass was disinterred from the carbonate substrate, following the published protocol (10). Lightly bored calcite chips with the external biomass lightly brushed free were prechilled on ice in 1.5-mL microcentrifuge tubes containing 300 μL PES-30 media for 30 min. These chips were then subjected to EDTA disinterment, but without continual brushing, as soon as boring “beds” of filaments became loose, the entire bed was transferred, using sterile fine-tipped forceps, to a 1.5-mL microcentrifuge tube containing chilled fresh PES-30 medium. Beds were gently washed twice without centrifugation and resuspended in PES-30 containing 20 μM BD-TH and incubated for 3 h, at room temperature, in the dark. Samples were carefully washed twice without centrifugation, and fresh medium was added before imaging.

BCECF-AM Ratiometric Imaging.

To assess intracellular pH, we used the ratiometric intracellular pH indicator BCECF-AM. This indicator has a pH-independent (isobestic) excitation point (∼445 nm) and a pH-dependent excitation point (488–496 nm). The closest laser excitation wavelength to the isobestic point in our laser scanning confocal fluorescence microscopy was 458 nm, which, although not optimal, was sufficient to detect changes with significance. We attempted to create a pH calibration curve using nigericin/valinomycin to obtain absolute pH values; however, the ionophores showed little effect over a large range of tested concentrations in M. testarum. However, we were able to obtain a good standard curve using N. punctiforme ATCC 21933, which closely followed an exponential regression between pH 6 and 9.5. Before loading Nostoc with dye, cells were incubated in BG11° containing 10 mM EDTA, spun, washed, and resuspended in fresh BG11° containing 25 μM BCECF-AM and incubated for 30 min at room temperature in the dark. These cells were then transferred into the various pH-calibrated buffers containing 50 μM nigericin and 5 μM valinomycin and imaged after a 30 min incubation. Intracellular BCECF values were determined by isolating only pixels contributing to intracellular signal by using image masking. Masking was done similarly to in Corvini et al.; however, image thresholding was performed using Li’s iterative algorithm for minimum cross entropy thresholding, which is integrated into ImageJ (36, 37). Each channel stack was multiplied by the masks, and then the projected (summed projections) 496-nm image was divided by the 458-nm channel in ImageJ (pixel by pixel) to obtain a 32-bit ratiometric image. This final image represents the relative intracellular pH values for the experiment. Each cell was assigned as a region of interest, and average cellular ratio values were plotted against respective pH values. The resulting regression curve was used to approximate the intracellular pH in Mastigocoleus.