Diffusion-controlled metabolism for long-term survival of single isolated microorganisms trapped within ice crystals

Abstract

Two known habitats for microbial metabolism in ice are surfaces of mineral grains and liquid veins along three-grain boundaries. We propose a third, more general, habitat in which a microbe frozen in ice can metabolize by redox reactions with dissolved small molecules such as CO₂, O₂, N₂, CO, and CH₄ diffusing through the ice lattice. We show that there is an adequate supply of diffusing molecules throughout deep glacial ice to sustain metabolism for >10⁵ yr. Using scanning fluorimetry to map proteins (a proxy for cells) and F420 (a proxy for methanogens) in ice cores, we find isolated spikes of fluorescence with intensity consistent with as few as one microbial cell in a volume of 0.16 μl with the protein mapper and in 1.9 μl with the methanogen mapper. With such precise localization, it should be possible to extract single cells for molecular identification.

Until recently it was not understood how microorganisms trapped in ice obtain the water, energy via redox reactions, and bioelements such as carbon required for life. Price (1) provided part of the answer. He developed a quantitative picture of how micrometer-sized bacteria and archaea could be swept into veins in a growing ice cap, along with chemical impurities of aeolian origin that are insoluble in the ice. In equilibrium the freezing point of an aqueous solution is lower than that of pure water. In polycrystalline ice the solute ions concentrate at the triple junctions of grains, leading to a vein structure. Price calculated the molarity and the vein diameter as a function of freezing point depression for the concentration of the major solute ions in ice. Vein diameters range from ≈1 μm at −50°C to ≈10 μm at −4°C in glacial ice. Two disadvantages of this habitat are that large microbes do not fit into veins, and those small enough to fit must adapt to a harsh chemical environment such as sulfuric acid. Nevertheless, for those adapted to such an environment, redox reactions of impurities in the veins yield enough energy to sustain a substantial microbial population.

Junge et al. showed that many microorganisms in sea ice occupy brine channels (2) and that many in lake ice occupy veins (3). Mader et al. (4) showed that both bacteria and fluorescent beads added to water used to make ice are rejected from the solid phase and incorporated into liquid veins, provided that they are small enough to fit, whereas beads larger than the vein diameter are frozen into solid ice.

Tung et al. (5) proposed a second icy habitat, which is afforded by surfaces of mineral grains around which the freezing point of the aqueous solution is depressed within the “hydration distance.” Microbes attached to minerals extract energy in redox reactions with ions in the mineral grain. Wettlaufer (6) accounted for equilibrium undercooling, T_m − T, as a function of thickness of the “unfrozen” water as the sum of four terms: depression of the freezing point of an ionic solution, an attractive van der Waals contribution, a Debye–Hückel contribution due to charges on the mineral surface, and a Gibbs–Thomson contribution, which is a function of curvature of the walls of any pores present. In the case of clay grains, the thickness of the unfrozen layer decreases from tens of nanometers at temperatures near 0°C to ≈0.3 nm at −80°C (7). Experiments (8, 9) have shown that the unfrozen water still has considerable mobility at these low temperatures. Price and Sowers (10) have presented evidence for microbial metabolism at temperatures as low as −40°C (water activity a_w = 0.72).

Up to 95% of the 10⁸ to 10¹⁰ cells per cm³ in the basal ice cored from a depth of 3,041–3,054 m in the Greenland Ice Sheet Project 2 (GISP2) are attached to clay grains (5). At four depths, Miteva et al.^† found that up to ≈85% were alive. Reasoning from their observed linear correlation of the number of cells per grain with grain circumference and from the factor of ≈10⁴ higher electronic conductivity parallel to the basal plane than perpendicular to it in Fe-rich clays (11), Tung et al. (5) inferred that the majority of the attached cells were Fe reducers metabolizing by electron shuttling. With this mechanism, they were able to explain how Fe reducers could reduce nearly 100% of all Fe³⁺ in clay grains. With epifluorescence microscopy of F420 [an autofluorescing coenzyme that is accepted as a unique signature of methanogens (12)], they determined that ≈2.4% of the cells in the basal ice were methanogens.

Need for a Third Microbial Habitat in Ice

Although the veins and mineral surfaces provide habitats for many microbes, observations suggest that these are unlikely to be the only locations where life endures in glacial ice. Numerous papers report the identification of microbes of diverse taxa, including nonextremophiles in ice (13–22). Eukarya up to ≈10² μm in size, some of which are viable, have been found in glacial ice (23–27). Because glacial ice is coldest at or near the top, the veins formed during grain growth and recrystallization will be the smallest in diameter and have the highest ionic concentration there. One might expect then that only extremophiles and the smallest microbes would be able to survive this harsh environment.

Baker et al. (28) and Barnes and Wolff (29) used scanning electron microscopy with energy-dispersive x-ray spectroscopy to map the location and composition of soluble impurities in glacial ice. Barnes and Wolff detected veins only at depths where the bulk concentration of ions (mainly sulfate) was greater than ≈1.6 μM and where the bulk ice was acidic. They suggested that vein networks do not form unless the acidic impurities are sufficient to coat all grain boundaries with at least a monolayer. Baker et al. (28) found that veins are present only in interglacial-stage ice, where the bulk pH is acidic. Thus, veins as a microbial habitat may be absent in ice with low acidity or large grain size.

Experimental evidence for the presence of both aerobes and anaerobes at the same depth in glacial ice provides a further constraint on habitat. Sheridan et al. (13) and Miteva et al. (14) identified a rich variety of both aerobes and anaerobes at the same depth, 3,043 m, in a sample of GISP2 ice. Using scanning fluorimetry to scan GISP2 ice cores at the National Ice Core Laboratory (NICL), we recently found that anomalously high levels of both δ¹⁸O_air (30) and CH₄ (31) at the same depth, 2,672 m, corresponded to excess microbial concentrations localized within a 1-cm³ ice volume^‡. We concluded that these gas anomalies are the waste products of both aerobic respiration and methanogenic metabolism within the same community. Because methanogens are among the most strictly anaerobic microorganisms and will not grow or make CH₄ in the presence of oxygen (32), this suggests that obligate aerobes and methanogens must have access to separate, isolated microenvironments within the same ice, rather than coexist in veins.

In view of all these limitations, we propose a third icy habitat that can accommodate microbial members of all three domains, of any size, at all depths, independent of oxygen content in the ice. In the next section, it will become clear that the habitat is so confining that it cannot permit movement or growth, but only survival.

Habitat 3: Interior of an Ice Crystal Far from Veins and Grain Boundaries

It is not obvious that a habitat in solid ice would permit survival, to say nothing of movement or growth. To set the stage, we model the behavior of a single methanogen frozen into the interior of an ice crystal at a temperature −10°C. For GISP2 ice, this corresponds to a depth ≈3,000 m and pressure of ≈30 MPa. In equilibrium at −10°C, an isolated cell will be coated with ≈1 nm of unfrozen water (33), which provides easy access of nutrient molecules to the cell membrane. We take the mass per cell to be 38 fg for cells of the average size measured by examining SEM micrographs of cells from the deep GISP2 ice core (5). That mass corresponds to ≈10⁹ carbon atoms and a radius of ≈0.21 μm for a cell of spherical shape. Price and Sowers (10) found that the typical rate of carbon turnover for survival metabolism in a microbial community at −10°C is 2 × 10⁻⁶ g C·(g C)⁻¹·yr⁻¹ (in which g C is grams of carbon), which would require only ≈1,900 carbon atoms to be converted from CO₂ to CH₄ per cell per year to repair molecular degeneration, assuming the reaction

A methanogen in ice could thus be a sink for dissolved H₂ and CO₂ molecules and a source of CH₄ and H₂O molecules. The CH₄ would be trapped in the ice but could diffuse short distances.

To determine whether the reaction in 1 can be sustained in ice at the rate of 1,900 molecules per year per cell, we need to know the equilibrium concentrations of CO₂ and H₂ dissolved in ice and the diffusion coefficients of CO₂, H₂, CH₄, and H₂O in ice at −10°C. Later we will generalize the discussion to metabolism of cells of larger size, at other temperatures, for other redox reactions, including aerobes and anaerobes. Supporting information (SI) Table 1 gives the measured or estimated diffusion coefficient, D, within an ice crystal lattice (as distinct from diffusion along veins and grain boundaries, which is much faster) for a number of atoms and molecules. All of the gas molecules relevant to climate studies have values of D ≤ 10⁻¹⁴ m²·s⁻¹, consistent with survival of observed sharp changes of concentration on timescales >10⁵ yr. SI Table 2 gives equilibrium concentrations of dissolved molecules in the ice lattice, under the assumption that the gases in the ice are in equilibrium with air bubbles in the ice.

As a result of snow densification and pore closures in glacial ice, ≈10% of the firn volume becomes trapped in ice as air bubbles. At sufficiently high pressures, the bubbles become unstable against transformation into the air-hydrate clathrate phase (34). Lipenkov (35) reviewed the measurements of sizes and concentrations of air bubbles and of air-hydrate crystals as a function of depth in the Vostok ice core, and Pauer et al. (36) did the same for air in the Greenland Ice Core Project (GRIP) core. Due to the long duration for nucleation and growth of an air-hydrate crystal at a bubble, air-hydrate crystals and bubbles coexist at depths ≈500 to ≈1,250 m in Vostok ice and at ≈650 to 1,300 m in GRIP ice. Lipenkov (35) concluded that the conversion efficiency is one air-hydrate crystal per air bubble. For depths shallower than the air-hydrate phase boundary, the equilibrium molar fraction, C_e, of gas in the ice lattice is proportional to pressure, as given in SI Table 2. For greater depths, C_e is approximately constant and equal to the concentration at nucleation (37).

Diffusion of gases dissolved in the ice lattice has been experimentally observed in a number of situations, including: diffusion of gas from small bubbles to larger bubbles, diffusion from bubbles to air-hydrate crystals, and diffusion from small hydrate crystals to large ones (37). Further, each of these observed processes can be understood as the result of a slowly diffusing gas seeking the state of lowest free energy. From measurements of relative sizes of air-hydrate crystals at typical separations of ≈1 mm and at depths 1,350 to 2,040 m, Uchida et al. (37) obtained an average value of the product of the diffusion rate, D, and the equilibrium molar fraction, C_e, as DC_e ≈ 10⁻¹⁹ m²·s⁻¹ at a temperature of −40°C. This is consistent with the values from SI Tables 1 and 2 for oxygen and nitrogen and implies a gas mobility of several tens of centimeters per 10⁵ yr.

Time for Replenishment of Molecules Reacted During Metabolism of an Isolated Cell

We now show that an isolated microbial cell of volume V_cell in an ice lattice can metabolize at the rate μ(T) [in g C·(g C)⁻¹·yr⁻¹] necessary for survival (10) by using suitable impurities that are present in equilibrium concentration C_e (in mole fraction) and that diffuse through the ice with diffusion coefficient D(T). Diffusion is governed by the equation

By enforcing spherical symmetry and matching boundary conditions, an exact steady state solution to this can be found as

with r_cell denoting the radius of the cell. It follows that the number of molecules entering a spherical cell per unit time is 4πD(T)C_er_cellρ_ice/MW_ice, where ρ_ice and MW_ice are the density and molecular weight of ice. Equating this to the metabolic requirements of the cell, we find that for a given metabolic rate, diffusion rate, and equilibrium concentration, the largest sustainable cell is given by

Consider metabolism (1) at −10°C by methanogens at a rate μ(−10°) = 2 × 10⁻⁶ g C·(g C)⁻¹·yr⁻¹ (10) at a 3,000-m depth in GISP2 ice. For a typical 38-fg cell (5), this corresponds to a consumption of only 1,900 C atoms per year. As 3,000 m is far below the clathrate transition zone, the equilibrium concentration is determined by the clathrate dissociation pressure (6.2 MPa at −31°C). From SI Table 1, D(CO₂) = 10⁻¹⁴ m²·s⁻¹, C_e(CO₂) = 8 × 10⁻¹⁰ moles·mole_ice⁻¹·MPa⁻¹ = 5 × 10⁻⁹ moles·mole_ice⁻¹, and D(H₂) = 1 × 10⁻⁸ m²·s⁻¹. For hydrogen, we note that air containing ≈550 ppb of H₂ has too low a concentration to reach the equilibrium with bubbles described in SI Table 2, meaning that essentially all of the H₂ available in air becomes dissolved in the ice. Given that glacial ice at the elevation of GISP2 forms with ≈9% air by volume (38), this gives C(H₂) = 4 × 10⁻¹¹. From SI Table 1 we see that the products H₂O and CH₄ will diffuse out of the shell at rates comparable to that of CO₂, so that the waste products do not build up around the cell. From these values we see that methane production is limited by the abundance of CO₂ and from Eq. 4 that cells with a diameter as large as 76 μm will receive sufficient CO₂ to sustain survival metabolism. Hence, this accommodates archaea, bacteria, and even quite large eukaryotes.

At a depth of 100 m (1 MPa) where the temperature is much lower (−32°C), the metabolic rate is μ(−32°) = 5 × 10⁻⁸ g C·(g C)⁻¹·yr⁻¹ (10) or a turnover rate of only 48 carbon atoms per year for the typical cell. We get from SI Tables 1 and 2 that D(CO₂) = 10⁻¹⁵ m²·s⁻¹ and C_e(CO₂) = 8 × 10⁻¹⁰ moles·mole_ice⁻¹, which accommodates cells up to 60 μm in diameter. In this example, the lower diffusion rate at −32° is largely compensated for by the lower metabolic requirements.

For other taxa, metabolism can proceed at the rate given by Price and Sowers (10), provided diffusion is rapid enough to maintain the supply of reactants. Examples include aerobic respiration (e.g., CH₄ + 2O₂ → 2CO₂ + 2H₂O), nitrogen fixation (N₂ + 6H⁺ + 6e⁻ → 2NH₃), and carbon monoxide oxidation (e.g., O₂ + CO → CO₂), all of which involve small molecules with values of D not smaller than D(CO₂) (SI Table 1). N₂, O₂, CH₄, and CO are relatively abundant in the atmosphere: 565 ± 200 ppbV for CH₄ (39) and 100 ± 50 ppbV for CO (40). All of them form clathrate hydrates with known phase diagrams, and their solubilities in ice as a function of temperature can be estimated from their dissociation pressures (41–43). We conclude that the above metabolic processes should not be restricted by the values of D or C_e.

As a final example, we consider aerobic consumption of CH₄ at 3,000-m depth and −10°C. The maximum size cell that can respire aerobically at 2 × 10⁻⁶ g C·(g C)⁻¹·yr⁻¹ via a reaction such as CH₄ + 2O₂ → CO₂ + 2H₂O can be estimated with D assumed to be 10⁻¹⁴ m²·s⁻¹ and C_e for methane assumed to be 4 × 10⁻¹¹ moles·mole_ice⁻¹ (limited by atmospheric abundance). This allows a spherical cell with diameter as large as 5 μm to metabolize at a survival rate at −10°C. Even if D were as low as 10⁻¹⁵ m²·s⁻¹, a concentration of methane near a cell 1.6 μm in size could sustain survival metabolism.

Sustainability of Intraglacial Metabolism

The last factor to consider is whether metabolism by diffusion through glacial ice can be sustained for the entire age of the ice. For a given metabolic rate and concentration of microbes, we can approximate the time required to deplete the reactants as 2C_eρ_iceMW_carbon [n_cell m_cell μ(T) A MW_ice]⁻¹, where n_cell is the number density of cells, m_cell is the mean mass per cell, and A is the number of moles of reactant per mole of carbon metabolized. For a lifetime τ, this can be recast to note that the sustainable concentration of cells is approximately

For τ ≈ 100 kyr, cells of mass 38 fg, a metabolic rate of 10⁻⁶ g C·(g C)⁻¹·yr⁻¹ at −10°C (10), a concentration of ≈4 × 10⁻¹¹ moles·mole_ice⁻¹ (the case for hydrogen and methane), and A = 4 from reaction 1, we find that the supportable population of methanogenic archaea by diffusion to isolated cells is ≈2,000 cm⁻³. An additional ≈8,000 small cells per cm³ are supportable via aerobic methane consumption. These estimates are conservative in that they ignore the period spent at lower temperature, which would increase the sustainable population size, due to the lower metabolic rate.

Fluorimetric Evidence for Isolated Single Microbes

We used two new scanning fluorimeters to search for evidence for the existence of isolated microbes in ice cores stored at −36°C at NICL. To map microbial cells via their protein autofluorescence (dominated by protein-bound tryptophan), we used a Photon Systems (Covina, California) TUCS (Targeted UV Chemical Sensor) fluorimeter with a 224-nm laser source. The cylindrical volume sampled has a radius ≈100 μm and depth ≈0.5 cm governed by the laser beam diameter, depth of focus, and acceptance of the six phototubes. To map the coenzyme F420 in methanogens, we used a compact 404-nm laser fluorimeter of our own design, which samples a volume with radius ≈350 μm and depth 0.5 cm. The two instruments take readings at ≈300-μm depth intervals.

Fig. 1 shows protein fluorescence as a function of depth in several regions of West Antarctic Ice Sheet (WAIS) Divide and GISP2 ice cores. Red points have spectra consistent with protein fluorescence; blue points have a different spectral shape. Four of the high points are labeled by numbers indicating the number of cells with an average mass of 38 fg within an illuminated spot. The high fluorescence intensity in the spot suggests that the laser beam passed through a vein containing a high concentration of microbes. At three depths where several points with high intensities fall close to a vertical dotted line, the laser beam may have traversed a vein at a small angle, resulting in excitation of protein fluorescence several times. Several points marked with an arrow indicate signals with intensity consistent with fluorescence of single microbes. Points with higher and lower intensity are likely due to microbes with higher and lower protein concentrations, surviving in the ice lattice as isolated cells. Such points provide our best evidence that isolated cells not associated with veins exist in glacial ice.

Fig. 1.

Examples of protein fluorescence intensity vs. depth obtained with the TUCS in scans of ice cores from WAIS Divide (Antarctica) and GISP2 (Greenland). Red points have spectra consistent with protein fluorescence.

Fig. 2 shows the F420 fluorescence intensities in several regions of WAIS Divide and GISP2 ice cores. Arrows indicate several spikes with intensity corresponding to that of a single 38-fg methanogenic cell.

Fig. 2.

Examples of F420 fluorescence intensity vs. depth obtained with the F420-fluorimeter.

Discussion

We have developed a quantitative underpinning for a habitat for isolated microbes in solid ice and presented evidence, via scanning fluorimetry, that microbes occupy this habitat. We showed that at all depths there is an adequate concentration of small molecules of atmospheric origin dissolved in the ice lattice and available for metabolism by isolated microbes. At shallow depths, small air bubbles shrink by diffusion of gases from their surfaces to larger bubbles, driven by the Gibbs–Thomson effect. At depths where all air bubbles have converted to air-hydrate crystals, their rate of change of size with depth (i.e., time) provides an adequate supply of dissolved gases to meet the needs of the very slow metabolism expected of microbes frozen into the ice, even for a time >10⁵ yr.

Using our scanning fluorimeters, we mapped microbial distributions and methanogen distributions at 300-μm depth intervals in a number of ice cores. The data suggest that microbes in both Greenland and Antarctic ice have a broad distribution of cell sizes. The results support our calculations that microbes up to at least 5 μm in diameter survive in the ice lattice, even if far from veins and mineral surfaces.

The proposed habitat avoids problems associated with the vein habitat, including the toxicity of reactants in veins, their narrow width, their lack of continuity throughout glacial ice, and the presence or absence of free oxygen in veins. With an equilibrium concentration of 2.4 × 10⁻⁷ mole fraction O₂ dissolved in the lattice, both aerobes and anaerobes should survive as isolated cells in the lattice. Because this might not be obvious, we elaborate: An isolated aerobe can respire by diffusion of O₂ and CH₄ molecules from the ice lattice to its cell membrane. Low levels of the enzymes catalase and superoxide dismutase have been detected in methanogens (44, 45), and these may serve as adequate protection against the toxic products of oxygen reduction at that low concentration of O₂.

One might ask how nonpsychrophiles can survive when incorporated into the ice lattice at an ambient temperature as low as −55°C (in East Antarctic ice) and at a hydrostatic pressure that may exceed ≈30 MPa at the bottom of a growing ice sheet. Smith et al. (46) showed that after 2 months of exposure in the Southern Ocean, several taxa of mesophiles including Escherichia coli were able to grow at −1.8°C and 34.5 ppt salinity and were no longer able to form colonies at 37°C (their previous optimal growth temperature). Sharma et al. (47) discovered that E. coli and Shewanella oneidensis remain viable and motile in liquid veins in ice for more than 1 month at pressures up to 1,600 MPa. These and other experiments show that nonpsychrophiles deposited onto a growing ice cap may adapt to both low temperatures and high pressures.

The requirement that the molecules have sufficiently high diffusivity and equilibrium concentration in solid ice to maintain metabolism for more than 10⁵ yr rules out polyatomic molecules with more than about five atoms: Methane will work but acetate has too low a diffusivity. One might ask whether microbes that prefer large molecules for their metabolism and that grow best in the presence of catalysts can live isolated in ice with access only to small molecules. Faced with the necessity of doing without, such microbes may express genes that allow them to use an alternative source such as methane + oxygen.

We have shown that values of D ≈ 10⁻¹⁵ to 10⁻¹⁴ m²·s⁻¹ are sufficiently high to sustain metabolism at all depths in glacial ice. Conversely, using the expression x ≈ (Dt)^1/2 to estimate diffusion distance in time t, and values of D in SI Table 1, we see that initially sharp profiles in ice of gases used in studies of climate change (CO₂, CH₄, N₂O, HDO, and H₂¹⁸O) will be blurred by no more than a few cm on a timescale of 10⁵ yr.

It is worth reiterating that in solid ice, microbial cells survive at a metabolic rate just great enough to repair macromolecular damage, which is ≈6 orders of magnitude lower than that necessary for exponential growth and mobility (10). Thus, the problem of carving out room for cell division in the ice does not arise.

It may be possible to exploit the ability of the TUCS scanning fluorimeter to localize single cells within its active volume of 160 nl. Westphal et al. (48) have developed a method to extract micrometer-sized interplanetary dust grains from an aerogel target on the National Aeronautics and Space Administration Stardust mission by using a “nanomanipulator” to cut out a small portion of the aerogel containing the grain. Such a device could be used in a cold room to cut out the volume of ice associated with a given fluorimeter signal. Then, under sterile conditions, the ice could be melted, and a single-cell genetic analysis could be done (49) on one or more cells located with the fluorimeter.

Materials and Methods

We assume that intensities of protein and F420 fluorescence per cell are independent of time for living microbes in ice, on the grounds that their survival metabolism is sufficient to repair macromolecular damage (10).

With the TUCS we did ground truth calibrations at depths 2,953.6, 2,953.95, 3,000, 3,018, and 3,035.88 m in GISP2 by comparing fluorescence intensity, I_TUCS, with counts of stained cells (50). We obtained the following conversion factors: Assuming an average cell mass of 38 fg (5), the background level for the TUCS fluorescence signal corresponds to ≈0.5 to ≈1.4 cell; the number of cells detected in the active cylindrical volume of depth 0.5 cm is 243 × I_TUCS; and the concentration is C_cells = 1.52 × 10⁶ cm⁻³ × I_TUCS. Fig. 3(Right) shows five-channel fluorescence spectra of E. coli (green curve), Bacillus spores (blue), kaolinite grains (pink), and volcanic ash (yellow) measured with the TUCS. Protein fluorescence is characterized by having the highest intensity at 320 and 340 nm. Fluorescence of mineral grains is several orders of magnitude weaker than that of microbial cells and usually has maximum intensity at wavelengths >380 nm. The curves for kaolinite and ash are scaled up by a factor 100 for visibility. Fig. 3 (Left) shows TUCS intensity at 300-μm depth intervals at a depth of 2,953.9 m in a 1-m-long GISP2 ice core chosen for calibration purposes because Tung et al. (50) had found a huge excess of microbial cells including methanogens in a 20-ml sample from that depth. The full spectra for every point are stored on disk. The points in red are for spectra with shapes that correspond to protein fluorescence. The points in blue have different emission spectra, probably due to mineral dust grains in the ice.

Fig. 3.

Calibration of TUCS to microbes in ice. (Left) Intensity of fluorescence in the TUCS along a 1-m section of GISP2 ice. Points in red indicate protein fluorescence; points in blue have spectra inconsistent with proteins and are likely inorganic. (Right) Spectra for measurements in the boxed region compared with spectra for lab specimens for bacteria and minerals.

Fig. 4 shows our F420 fluorimeter in position to scan an ice core. It uses a 404-nm laser to excite F420 fluorescence in a volume 1.9 × 10⁻³ cm³. To calibrate this instrument, we did scans along the same regions of GISP2 ice cores with very high microbial concentrations where Tung et al. (50) counted methanogens with epifluorescence microscopy. We found that the background level corresponds to 0.15 to 0.3 methanogenic cell of mass 38 fg. The number of methanogens detected in the active volume is 1.94 × 10⁻³ × I_F420, and the concentration is C_meth = 1.02 cm⁻³ × I_F420. We note that F420 fluoresces only if it is in an oxidized state and if its pH is greater than 6 (51). During metabolism, 80% of F420 is in the oxidized state (52). When the cell stops metabolizing, the amount of oxidized F420 decreases; thus, F420 fluorescence is an indicator of active methanogenesis. The concentration of F420 in a living methanogen depends not only on its cell mass but also to some extent on its species (53).

Fig. 4.

F420 fluorimeter with 404-nm laser (vertical black cylinder) surrounded by seven photon counters covered by band-pass filters, fixed at an adjustable height above an ice core on a translation stage.

Abbreviations

GISP2

Greenland Ice Sheet Project 2

WAIS

West Antarctic Ice Sheet.