Fluorescent Fingerprints of Endolithic Phototrophic Cyanobacteria Living within Halite Rocks in the Atacama Desert

Abstract

Halite deposits from the hyperarid zone of the Atacama Desert reveal the presence of endolithic microbial colonization dominated by cyanobacteria associated with heterotrophic bacteria and archaea. Using the λ-scan confocal laser scanning microscopy (CLSM) option, this study examines the autofluorescence emission spectra produced by single cyanobacterial cells found inside halite rocks and by their photosynthetic pigments. Photosynthetic pigments could be identified according to the shapes of the emission spectra and wavelengths of fluorescence peaks. According to their fluorescence fingerprints, three groups of cyanobacterial cells were identified within this natural extreme microhabitat: (i) cells producing a single fluorescence peak corresponding to the emission range of phycobiliproteins and chlorophyll a, (ii) cells producing two fluorescence peaks within the red and green signal ranges, and (iii) cells emitting only low-intensity fluorescence within the nonspecific green fluorescence signal range. Photosynthetic pigment fingerprints emerged as indicators of the preservation state or viability of the cells. These observations were supported by a cell plasma membrane integrity test based on Sytox Green DNA staining and by transmission electron microscopy ultrastructural observations of cyanobacterial cells.

INTRODUCTION

The Atacama Desert in northern Chile is the driest place on Earth. Its most arid regions lie between the Andes rain shadow and the Coastal Cordillera (1). Until recently, this hyperarid zone, or central depression, of the Atacama Desert was considered the dry limit for photosynthetic activity (2) and primary production. Indeed, only extremely low concentrations of microorganisms had been detected in the soils of this zone (3). However, despite the major constraint for microbial life in the desert being the scarcity of liquid water, an abundance of photosynthetic life, mainly cyanobacteria associated with heterotrophic bacteria and archaea, has recently been detected in the Atacama's hyperarid zone (4–6). This remarkable discovery was made inside the halite deposits that form part of the Neogene salt-encrusted playas of Atacama, also known as salares (7). Chroococcidiopsis-like cells were the only cyanobacteria found inside halite pinnacles, and phylogenetic studies revealed their close genetic affinity to the genus Halothece (5, 6). The presence of this endolithic community indicates that life has found a survival strategy in the hyperarid zone of Atacama, where all other colonization strategies have failed. These cells seem to be biologically adapted to conditions of high salinity, and their presence indicates a water source other than the area's practically nonexistent rainfall. Davila et al. (8) showed that water vapor condenses within the halite pinnacles at relative humidity (RH) levels that correspond to the deliquescence point of NaCl (RH = 75%). More recently, it was shown that halite endoliths could obtain liquid water through spontaneous capillary condensation at a relative humidity much lower than the deliquescence RH of NaCl (9).

All desert microorganisms undergo extended periods in a desiccated, metabolically inactive state in which individual cells are subjected to a variety of chemical and physiological stresses. These stresses lead to damage that cannot be repaired until metabolism restarts (10, 11). Besides desiccation, microbial communities inside halite crusts have to deal with conditions of extreme salinity (5, 6) and excess light, which have a direct effect on their physiology. In response to varying light conditions, photosynthetic microorganisms undergo structural, behavioral, physiological, and chemical modifications (12, 13), such as changing the quality and concentration of their light-harvesting pigments (12, 14, 15) and pigment degradation (11, 16, 17). This suggests that the state of photosynthetic pigments can be an indicator of cell viability.

The physiology of photosynthetic microorganisms can be investigated by examining their photosynthetic capacity. Because photosynthesis is such a rapid process and because it can only be indirectly inferred from measurements of related variables (e.g., oxygen and/or carbon dioxide), microscopy techniques used in multidisciplinary studies have emerged as powerful tools for the in vivo quantification of photosynthesis. Although these techniques are all noninvasive, each method has technical, practical, and physiological advantages and limitations to be considered (18, 19). Thus, some microscopy procedures can be used to detect physiological and biochemical changes in photosynthetic microorganisms, allowing the detection of fluorescence properties of photosynthetic pigments (17). Significant progress is constantly being made in detectors and computer technology, image analysis, visualization, laser sources, and optical technologies. These developments have allowed the noninvasive detection of photosynthetic pigment changes occurring in microorganisms in their natural environment (18, 20).

The present study was designed to detect autofluorescence emission spectra emitted by both single cells and the photosynthetic pigments of cyanobacteria living within halite rocks in the Atacama Desert. The fluorescent fingerprints obtained were used to identify cyanobacterial photosynthetic pigments and as a measure of cyanobacterial cell viability.

MATERIALS AND METHODS

Samples and study area.

This study compares the microbial colonization of halites from two different areas of the Atacama Desert (northern Chile): Yungay (24°5′53″S, 69°55′59″W) and Salar Grande (21°8′54″S, 70°1′04″W). Although both sampling areas occur in the hyperarid area of the Atacama Desert, their environmental conditions vary slightly (Table 1). The Yungay area shows a mean annual precipitation below 2 mm year⁻¹ and is considered the driest place on Earth (21). The area is 60 km from the coast and lies at an altitude of 962 m between the Coastal Cordillera to the west (1,000 to 3,000 m high) and the Domeyko Mountains to the east (about 4,000 m high). The other sampling site (Salar Grande) lies 300 km north of Yungay. However, its location close to the Pacific coast (8 km) and its lower altitude result in the region often experiencing the arrival of moist air and fog locally known as “camanchaca” (6, 22, 23).

TABLE 1.

Microclimate data recorded at sampling sites from May 2008 to May 2011

Site	Temp (°C)/yr				Relative humidity (%)/yr
	Mean	Maximum	Minimum	SDa	Mean	Maximum	Minimum	SD
Yungay	17.93	46.01	−8.2	11.30	34.52	76.55	2.40	21.38
Salar Grande	20.24	42.33	3.77	8.82	51.45	93.22	3.37	22.45

^aSD, standard deviation.

In both areas, the halite crust occurs as pinnacles that show characteristic irregular shapes formed by wind action and partial long-term dissolution and reprecipitation of evaporite deposits. At both sampling sites, air temperature (T) and RH were collected over a 3-year period (May 2008 to May 2011) using data loggers (Onset and HOBO Pro v2) as described by Wierzchos et al. (9). These RH and T sensors were placed close to the halite pinnacles 20 cm above the soil surface in the shade. Thus, the temperature recorded is a function of the air temperature and heat radiation from the nearby halite crust and soil.

According to detailed analyses of RH and T, photosynthetic active radiation and electrical conductivity sensor readings, as well as personal information obtained by two permanent workers at a remote water pump station 2 km from the Yungay area, there was no rainfall at the sampling sites from 2008 until May 2011. The RH and T data for the sampling sites are provided in Table 1.

The halite samples used for the present study were collected in January 2010 and May 2011 in expeditions to the sites mentioned above. Samples were stored dry in the dark at room temperature for no longer than 1 month until they were prepared for the different microscopy techniques. The day before preparation, they were left in a chamber under day/night light conditions at 75% RH to allow deliquescence and the adsorption of water by microbial cells.

Transmission electron microscopy (TEM).

Representative colonized halite layers were dissolved in a 20% aqueous NaCl solution and made up to an NaCl concentration of 5 M. After a short period (5 min) of precipitation of scarce mineral particles, the supernatant was centrifuged at 12,000 × g for 10 min. The precipitated microbial cells were fixed according to the protocol described by de los Ríos and Ascaso (24) with some modifications. In brief, these precipitates containing microbial cells were fixed in 3% glutaraldehyde in 5 M NaCl at room temperature for 3 h and then in 1% osmium tetroxide, dehydrated in a graded series of ethanol, and embedded in Spurr's resin. Poststained ultrathin sections were observed on a Zeiss EM910 transmission electron microscope equipped with a Gatan charge-coupled-device (CCD) camera.

Light, fluorescence, and confocal microscopy.

Halite samples taken from 3 to 5 mm below the crust surface containing pigmented microbial communities were scraped and dissolved in a 20% NaCl aqueous solution and made up to an NaCl concentration of 5 M. Following a short period (5 min) of precipitation of scarce mineral particles, the supernatant was centrifuged at 12,000 × g for 10 min. Pellets of microorganisms were resuspended in 20 μl of 5 M NaCl, and the cells were visualized by bright-field differential interference contrast (DIC) light microscopy using an AxioImager D1 Zeiss instrument equipped with a CCD color camera (AxioCam MRc; Zeiss) and a Plan-Apo 60×/1.4 Zeiss oil immersion objective. The same cell preparations were observed by fluorescence microscopy (FM) using specific filters for enhanced green fluorescent protein (EGFP) (Zeiss filter set 38; excitation/emission wavelengths, 450 to 490/500 to 550 nm) to visualize both the weak green autofluorescence of unidentified substances and specific intensive fluorescence of the Sytox Green (S-7020; Molecular Probes) dye. Also the rhodamine filter set (Zeiss filter set 20; excitation/emission wavelengths, 540 to 552/567 to 647 nm) was used to visualize the red autofluorescence of photosynthetic pigments.

To detect cyanobacterial cells with damaged membranes, the samples were stained using a specific-fluorescence Sytox Green dye. Some extent of cell membrane damage increases Sytox Green influx (10). This nucleic acid stain was used according to the method of Wierzchos et al. (25). The original solution containing 5 mM Sytox Green in anhydrous dimethyl sulfoxide (DMSO) was diluted to 1:100 in water and added to the suspension of halite-extracted microorganisms. The cells were stained for 10 min at room temperature and then examined by FM using a specific filter for EGFP (Sytox Green signal) using Plan-Apo 60×/1.4 and 100×/1.4 Zeiss oil immersion objectives.

Autofluorescence (green and red signals) was also visualized using a Leica TCS-SP5 confocal laser scanning microscope (CLSM) (Leica Microsystems Heidelberg GmbH, Mannheim, Germany) and an ×63 (1.4-numerical-aperture [NA]) Plan Apochromat oil immersion objective. Red autofluorescence was viewed in the red channel (640- to 785-nm emission) using a 561-nm laser diode, and green autofluorescence was observed in the green channel (495- to 560-nm emission) using a 488-nm line from an Ar laser, respectively. The samples were mounted on Mat-Teck culture dishes (Mat-Teck Corp., Ashland, MA, USA). Optical sections were acquired in x-y planes every 0.3 μm along the optical axis with a 1-Airy-unit confocal pinhole. Different projections were generated by the Leica LAS AF software and the Imaris software package, version 2.7 (Bitplane AG, Zürich, Switzerland) for three-dimensional (3D) reconstructions of cell aggregates.

Autofluorescence intensity was determined as an indicator of the integrity of the photosynthetic apparatus, as described by Billi et al. (10).

The emission spectra of cyanobacterial pigments were obtained using a wavelength λ-scan function of the CLSM based on a fluorescence method that determines the complete spectral distribution of the fluorescence signals emitted (20). Images were acquired with the same CLSM and objective. Series of images (xyλ) were taken to determine the emission spectra of the samples and to establish peaks. Photosynthetic pigments and other unknown autofluorescent molecules were excited with a 488-mm line of an argon laser. Fluorescence emission was captured in 10-nm bandwidth increments (lambda step size = 5 nm) in the range from 495 nm to 780 nm. A region of interest (ROI) in the thylakoid area was defined to determine the mean fluorescence intensity (MFI) in relation to the emission wavelength. A set of 20 ROIs of 1 μm² was used to analyze the mean fluorescence intensity and peak emission range of the samples. Fluorescence measurements were expressed in arbitrary units (AU).

RESULTS

Cyanobacterial cell ultrastructure.

Microbial communities inhabiting halite rocks in the Atacama Desert are mainly comprised of one cyanobacterial morphotype accompanied by heterotrophic bacteria and archaea (5, 6). The TEM images in Fig. 1 show the different types of aggregates containing cyanobacterial cells observed. Some of the round multicellular aggregates were composed of several cyanobacterial cells (Fig. 1a and b). Other aggregates showed a linear organization of phototrophic cells (Fig. 1c). The in situ visualization of this cryptoendolithic microbial ecosystem using low-temperature scanning electron microscopy, as well as FM and CLSM, suggests that the given aggregate type often depends on the shape of the pore spaces among the halite crystals occupied by microorganisms. Cells within the aggregates divide by binary fission to give rise to polygonal structures consisting of cyanobacterial cells embedded within an extracellular polymeric substance (EPS). In the TEM images, this can be seen as a nanoporous network surrounding the cyanobacterial cytoplasm (Fig. 1, asterisks). Cyanobacterial aggregates appear to be enveloped by an electron-dense fibrous outer layer, and in some cases (Fig. 1c), the cytoplasm of single cyanobacterial cells is also enveloped by this electron-dense fibrous layer (Fig. 1, open arrows). In many of the cells observed by TEM, the well-developed thylakoid system showed parallel membranes in interthylakoid spaces (e.g., Fig. 1b, white arrows). However, some cells revealed signs of senescence at the ultrastructural level (Fig. 1, black arrows). In these cyanobacterial cells, the thylakoid structure becomes disorganized and/or deteriorated, clearly indicating the degradation of cell structure, which likely disrupts photosynthetic activity and metabolism. Albeit indirectly, senescence is by far the leading cause of cell viability loss. Note that these senescent cells can be found within the same aggregates where cells with undisturbed ultrastructural elements, or potentially viable cells, are also present. The space around the cyanobacterial aggregates was filled with the remains of microbial cells (Fig. 1b). The cells of heterotrophic bacteria and archaea were frequently observed attached to the outer sheaths of cyanobacterial aggregates (Fig. 1a), or in some cases, isolated bacterial cells were found just beneath the outer layers of the cyanobacterial aggregates (Fig. 1c, black arrowhead).

FIG 1.

TEM images of cryptoendolithic microorganisms found in halites. (a) Round multicellular aggregate composed of cyanobacteria and heterotrophic bacteria (and/or archaea) (black arrowheads) adhering to the outer electron-dense multilayered structure (open arrows [also in panels b and c]) enveloping the aggregate. Note the extracellular polymeric substances surrounding cyanobacterial cells (asterisks [also in panels b and c]). (b) Aggregate with four cyanobacterial cells at different stages of senescence. The white arrows point to a cell with well-organized parallel thylakoids, the black arrows indicate cells with a high level of thylakoid disorganization, and the open arrowheads point to the remains of dead microorganisms. (c) Aggregate with linearly organized phototrophic cells, one of which (black arrow) shows a high level of thylakoid disorganization.

Red and green autofluorescence from cyanobacterial cells versus membrane integrity.

When examined by FM, the cyanobacterial cells forming the multicellular aggregates showed two distinct ranges of autofluorescence emission signal. Some of the cells emitted autofluorescence in the red signal range (detected using the rhodamine filter set), which corresponds to photosynthetic pigment autofluorescence (PAF). Other cells emitted less intense nonspecific broad-spectrum autofluorescence in the green signal range (GAF), detected using the EGFP filter set. Both kinds of emission signal were observed even among cyanobacterial cells within the same aggregate, as shown in Fig. 2a and b. The GAF signal observed suggests that some of the cyanobacterial cells contained degraded photosynthetic pigments. In GAF signal-emitting cells, autofluorescence was generally distributed evenly across the cell cytoplasm. However, the PAF signal showed a heterogeneous pattern in the cell cytoplasm, supposedly reflecting the positions of cyanobacterial thylakoids (data not shown).

FIG 2.

Green and red autofluorescence patterns observed for cyanobacterial cells within aggregates and plasma membrane-damaged cells in halite samples from Yungay. (a and b) Photosynthetic pigments emitting in the red (PAF) signal range and nonspecific green (GAF) signal range. (c and g) DIC images of cyanobacterial aggregates. (d and h) Photosynthetic pigment autofluorescence (the arrows indicate intense PAF signal). (e and i) Sytox Green signal (arrows) emitted by cells with damaged plasma membranes; cells with intact membranes are not labeled (black arrows in panels d, f, h, and j). (f and j) Merged color images of panels d plus e and h plus i, respectively.

The Sytox Green assay revealed the different viability states of cyanobacteria from the Yungay halites. The method identified cyanobacterial cells within an aggregate with damaged plasma membranes. The cells showed a Sytox Green (green-signal)-labeled DNA structure, indicating a loose membrane architecture (Fig. 2e, f, i, and j, white arrows). This signal was detected using the EGFP filter set and was much more intense than the weak GAF autofluorescence. Low-intensity or null red autofluorescence was also observed within the cells positive for Sytox Green. In contrast, cyanobacterial cells emitting a high-intensity PAF signal showed no Sytox Green labeling (Fig. 2d, f, h, and j, black arrows).

The same observations were made when Sytox Green was used on the endolithic microbial community from the Salar Grande halite. Figure 3 shows cyanobacterial aggregates and associated heterotrophic bacteria and archaea in DIC images after Sytox Green staining. This time, we were able to distinguish three types of aggregates containing cyanobacterial cells showing an intense PAF signal and a negative Sytox Green signal (Fig. 3b, aggregates with blue dotted outlines), cyanobacterial cells showing distinct PAF and Sytox Green signals (Fig. 3b, aggregates with yellow dotted outlines), and cyanobacterial cells showing a weak GAF signal yet a distinct Sytox Green signal (Fig. 3b, aggregates with white dotted outlines). We were also able to observe Sytox Green signals from dead bacteria and/or archaea outside the aggregates (Fig. 3b, white arrows).

FIG 3.

Endolithic microbial community found in the Salar Grande halites. (a) DIC image. (b) Fluorescence microscopy image. Aggregates containing cyanobacteria show an intense PAF signal (blue dotted outlines); aggregates containing cyanobacteria show both PAF and Sytox Green signals (yellow dotted outlines); aggregates containing cyanobacteria show a weak GAF signal but a strong Sytox Green signal (white dotted outlines). The arrows point to Sytox Green-stained dead bacteria and/or archaea in interaggregate spaces.

Fluorescence emission spectra of phototrophic cyanobacterial cells.

The lambda-scan confocal microscopy option records a series of individual images obtained using a defined emission fluorescence wavelength range. This procedure has been successfully used to assess the physiological state of photosynthetic microorganisms at the single-cell level (11, 17). We used the CLSM lambda-scan feature to characterize the fluorescence emitted by cyanobacterial photosynthetic pigments. When excited at a wavelength of 488 nm, cyanobacterial cells isolated from the halite samples showed three types of emissions: some emitted weak fluorescence within the green range, some within the red range, and others in both the red and green ranges (Fig. 4 to 6).

FIG 4.

Autofluorescence CLSM λ-scan images and corresponding emission spectra recorded in cyanobacteria isolated from halite (Yungay). (a) CLSM λ-scan image and spectral profile corresponding to nonspecific cell autofluorescence in the green region (GAF) in response to 488-nm laser excitation. Below is a plot of the MFI versus emission wavelengths of the cells (λ-scan emission maximum = 556.12 nm). (b) CLSM λ-scan image and spectral profile corresponding to the emission peaks of photosynthetic pigments (PAF) excited with the 488-nm laser. Below is a plot of the MFI versus emission wavelengths of the cells. Note the peak at 662.75 nm for PC and APC and the shoulder at 685.2 nm for chlorophyll a. The data from both spectra represent the MFIs (n = 15) ± standard errors (SE).

FIG 6.

Autofluorescence emission spectrum recorded for cyanobacteria isolated from halite (Yungay). The spectral profile reveals weak nonspecific autofluorescence in the green region and weak emission in the range of photosynthetic pigments. The cells were excited using a 488-nm laser. Plotted are MFI values (n = 15) ± SE against the emission wavelength of the cells.

Some emission spectra produced by cyanobacterial cells isolated from halites from both Yungay and Salar Grande featured a wide curve with a low maximum intensity at ca. 560 nm (Table 2 and Fig. 4a and 5a). These spectra were recorded in cells showing GAF-type emission that was confirmed by simultaneous (CLSM and FM) visualizations. In addition, many of the cyanobacterial cells from both locations exhibited a distinct emission peak between 657.6 and 662.7 nm (Table 2 and Fig. 4b and 5b). We consider that this high-intensity emission peak is produced by overlapping of the spectra of phycobiliprotein photosynthetic pigments whose characteristic emission peaks correspond to phycocyanin (PC) and allophycocyanin (APC) (20). Moreover, our spectra show an asymmetric slope with a small shoulder at ca. 680 nm, which could be the outcome of overlap of the characteristic emission spectra of phycobiliproteins and chlorophyll a (Chl a) (20). These spectra were recorded in cells showing PAF-type emission, as confirmed by simultaneous CLSM and FM visualizations.

TABLE 2.

Numerical values of maximum and mean fluorescence intensities of the λ-scan spectra emitted by different autofluorescence sources within cyanobacterial cells

Source of autofluorescence	Fluorescence intensity (±SE)
	Yungay		Salar Grande
	Maximum (nm)	Mean (AU)	Maximum (nm)	Mean (AU)
Pigments from degraded cells/GAF emission	556.12	43.33 ± 2.91	567.35	144.16 ± 8.65
Nondegraded photosynthetic pigments/PAF emission	662.75	151.40 ± 3.98	657.59	225.13 ± 19.06
Pigments from transition phase cells/GAF + PAF emissions	550.51	27.65 ± 2.57	NOa	NO
	662.75	36.2 ± 6.35

^aNO, not observed; SE, standard error.

FIG 5.

Autofluorescence CLSM λ-scan images and corresponding emission spectra recorded in cyanobacteria isolated from halite (Salar Grande). (a) CLSM λ-scan image and spectral profile corresponding to nonspecific cell autofluorescence in the green region (GAF) in response to 488-nm laser excitation. Below is a plot of the MFI versus emission wavelengths of the cells (λ-scan emission maximum = 567.35 nm). (b) CLSM λ-scan image and spectral profile corresponding to the emission peaks of photosynthetic pigments (PAF) excited with the 488-nm laser. Below is a plot of the MFI versus emission wavelengths of the cells. Note the peak at 657.59 nm for PC and APC and the shoulder at 679.6 nm for chlorophyll a. The data from both spectra represent MFIs (n = 17) ± SE.

Some cyanobacterial cells isolated from the Yungay halite showed an emission spectrum with two broad low-intensity peaks (Table 2 and Fig. 6). These peaks were recorded in cells showing both GAF- and PAF-type emissions, as confirmed by simultaneous CLSM and FM visualizations.

MFI and the half-bandwidth of the spectra differed for the GAF, PAF, and GAF-plus-PAF emission spectrum patterns. Generally speaking, MFI values were higher for PAF- than GAF-type spectra, though this difference was more evident for cyanobacterial cells isolated from the Salar Grande halite. The bandwidths of the two types of emission spectra also differed, being wider for the green emission than for the red.

Collectively, our TEM, FM, and CLSM, Sytox Green assay, and lambda-scan data indicate the following. (i) Some cyanobacterial cells emitted red fluorescence (PAF) derived only from photosynthetic pigments and were negative for the Sytox Green stain. These cells showed well-organized parallel thylakoids and preserved their ultrastructural integrity, as observed by TEM (Fig. 1a). The emission spectra likely corresponding to these cells showed a peak in the range characteristic for phycobiliproteins and chlorophyll a (Fig. 2f and j, black arrows; 3, blue outline; and 4b and 5b). These cells may be classed as intact and healthy. (ii) Other cyanobacterial cells were stained with Sytox Green but still showed weak PAF (Fig. 2f and j, white arrows, and 3b, yellow outline). These cells probably gave rise to the emission spectra showing two low-intensity peaks (Fig. 6). Although these cells emit fluorescence attributable to photosynthetic pigments, their cell integrity has been lost (Sytox Green penetrates the cell), so they are not vital. (iii) Finally, yet other cyanobacterial cells were stained with Sytox Green and emitted only a nonspecific GAF signal. These cells generated emission spectra with a peak in the green region (Fig. 3, white outline, 4a, and 5a). These cells could correspond to the cells showing extensive thylakoid disorganization, as observed by TEM. We interpret these cells as nonviable.

DISCUSSION

In extreme environments, photosynthetic microorganisms develop different strategies to survive the more hostile time intervals (26). For endoliths found within halite pinnacles in the Atacama Desert, intensive solar radiation, salinity, and lack of water (9) are the most important factors threatening their survival.

Cyanobacterial cell viability.

As the dominant group of photosynthetic halite colonizers, cyanobacteria appeared in different physiological/viability states within a single aggregate. Knowing the viability of photosynthetic organisms in their natural environment is essential to understanding the ecology of extreme-environment microbial ecosystems (27–30). Knowledge in this area is still far from complete.

The viability of photosynthetic organisms has so far been examined in different ways. In the microbial endolithic ecosystem examined here, prior CLSM/TEM studies confirmed the presence of viable cells (4) and characterized the ultrastructure and the integrity of cyanobacterial and bacterial cells (5). More recently, quantum yield fluorescence measurements (31) and carbon cycling rates, as indicated by the isotope contents (¹³C and ¹⁴C) of phospholipid fatty acids (PLFA) and glycolipid fatty acids (GLFA) (32), have been reported. Although a universally applicable viability determination method could be a utopian ideal, there is a clear need to determine how broadly current viability determination methods can be applied. Some studies have reported different results using the same viability determination method (33). Culture-independent viability indicators, such as those proposed here, might not be conclusive, especially in complex extreme microenvironments. Their use to investigate complex populations under natural conditions, however, has been encouraged (34). The Sytox Green method has been described as broadly applicable to photosynthetic microorganisms (10, 11). Multiparameter techniques have also clarified some questions regarding single-cell viability (35). In this study, as a measure of cell health, we assessed both cell membrane integrity and autofluorescence patterns produced by different emission wavelength ranges.

PAF and GAF fluorescence.

Emission spectra in both the red and green emission ranges were observed for the cyanobacteria isolated from halite pinnacles. The bandwidth of these emissions was wider in the green than in the red range. This is because red fluorescence is produced by photosynthetic pigments (mostly Chl a and phycobiliproteins) with defined emission peaks (11, 15, 17, 20). In contrast, GAF may be attributable to fluorescence emitted by different molecules with different widely overlapping emission spectra. In photosynthetic microorganisms, photosynthetic pigment autofluorescence is considered an indicator of cell viability (10, 35–37). A loss of pigment fluorescence (chlorosis) (38) has been correlated with decreased enzyme activity and increased membrane permeability (37) and may therefore be a useful indicator of senescence for any species of alga (37).

When red autofluorescence fades, a green nonspecific fluorescence observable at the same excitation wavelength may appear. Different species of microalgae and higher plants in different physiological states exhibit GAF of varying intensities (39). For algae or/and cyanobacteria, the use of this indicator has scarcely been explored, and the molecules responsible for GAF have not yet been clearly identified (39). Green autofluorescence can be induced by a variety of different molecules, such as flavonoids, flavins (e.g., reduced flavin adenine dinucleotide [FADH]) (40), cinnamic acids, betaxanthine, luciferin compounds (39), and pyridine nucleotides (e.g., NADH) (36, 41). NADH has often been associated with viability under UV excitation (350 to 360 nm) and may be found within most metabolically active prokaryotic and eukaryotic cells (41). The molecule FADH is a likely viability indicator candidate because it fluoresces at the same wavelength as these cyanobacteria (42). Thus, the bright-green autofluorescence of FADH at 530 nm (excited by a 488-nm laser) provides information on the oxidation state and metabolism of both bacteria and eukaryotes (43, 44).

A number of different factors affecting the efficiency of energy transfer from PC to Chl a will cause a drop in the fluorescence intensity of the photosynthetic pigments in each cell (45). Environmental factors, such as low-light stress (17), high-light stress, and nutrient stress, or death of part of the population due to ageing (10, 11, 37) have so far been identified. In general, increased concentrations of chlorophyll oxidation products have been observed in nutrient-depleted cells, but it is likely that specific chlorophyll transformation pathways vary between species (37). In the case of endolithic cyanobacteria living inside halite pinnacles, excessive solar radiation with a high UV fraction is known to produce significant stress. Effectively, in Yungay, these cyanobacteria produce considerable amounts of scytonemin, a UV-protective pigment (46, 47). Nitrogen starvation also leads to low levels of photosynthesis during nitrogen limitation (48, 49). Nitrogen stress affects energy transfer from PC and Chl a, and therefore, their spectral properties under both in vivo and in vitro conditions (45).

Another important factor causing photosynthetic-pigment damage could be the constant high salinity, long periods of dryness (9), and low RH values inside the halite pinnacles. Recently, Davila et al. (31) observed that the PSII of phototrophic cyanobacteria inhabiting halite pinnacles in the Yungay area was inactive below an RH of 60%. However, when the RH increased to above 70%, fluorescence appeared within minutes and stabilized at relatively low, but significantly positive, values. It is significant that activation of PSII in the halite cyanobacteria did not occur until liquid water was produced through deliquescence (8) and/or water vapor condensed within the nanopores of halite (9). Hence, it seems that photosynthesis in cyanobacteria can only be activated in the presence of liquid water (50). All these factors can cause different viability states of the cells. Automortality is closely associated with nonviability (51, 52). Nonviable cells are defined as cells that still have an intact cell shape but can no longer grow or divide. The loss of membrane integrity occurs in the later stages of automortality, resulting in the total disintegration of the cell (53, 54). However, before this final damage, cells can suffer different forms of injury to the cytoplasm's ultrastructural elements, as shown here by TEM (Fig. 1). Once this process begins, degradation of the photosynthetic pigments, in particular chlorophyll (55), and finally fragmentation of the genome lead to the final stage of autolysis. Billi et al. (10) were able to correlate DNA fragmentation and loss of red fluorescence in photosynthetic organisms. Our observations do not exclude the possibility that the nonspecific autofluorescence in the green region (GAF) observed in some cells is related to the degradation products of chlorophyll, as reported by Tang and Dobbs (39). GAF could also be the consequence of increased levels of denatured proteins, for example, following the rapid degradation of photosynthetic pigments (56). Thus, cells showing degradation of their photosynthetic pigments would have completely lost their photosynthetic capacity (53, 57). This means the nonviable cells will no longer show photosynthetic-pigment fluorescence, allowing them to be clearly identified in situ and at the single-cell level by means of CLSM λ-scanning, even in communities composed of numerous cells. However, the persistence of phycobiliprotein autofluorescence serves as a survival marker and has been related to genome stability, undamaged plasma membranes, and dehydrogenase activity upon rewetting (10). It therefore seems that viable and nonviable phototrophic endolithic cyanobacterial cells can be distinguished according to culture-independent viability indicators and fluorescent fingerprints supported by cell plasma membrane integrity testing and cell ultrastructural observations.

Conclusions.

The findings of this study indicate that the λ-scan option of CLSM microscopy can be used to determine the viability of cyanobacterial cells according to the autofluorescence of their photosynthetic pigments (PAF) and to a nonspecific GAF. Testing is performed in situ without disrupting the spatial integrity of structured microbial communities or denaturing their biomolecules. We propose the use of this method as an efficient tool for in vivo studies designed to address the cell physiology of unculturable endolithic phototrophic microorganisms inhabiting extreme environments.