Experimental Evidence for a Microbial Origin of Reduction Spots in Red Beds on Earth—and Mars?

1. Introduction

In mid-2024, during its exploration of the Neretva Vallis channel leading into Jezero crater, Mars, NASA’s Perseverance rover encountered bleached (ferric-oxide-depleted) spots in the ferric-oxide-pigmented sedimentary rocks of the Bright Angel formation (Hurowitz et al., 2025). Similar “reduction spots” are common in ferric-oxide-rich red beds on Earth and have long been considered a potential biosignature (e.g., Hofmann, 1990; Spinks et al., 2010).

Perseverance found two populations of spots. Brick-red mudstones of the Mount Spoonhead member contained pale greenish millimeter-scale halos around mottled dark cores. Pink/tan mudstones of the Cheyava Falls member contained organic matter and pale millimeter-scale “leopard spots” depleted in ferric iron, enriched in ferrous iron, and enriched in sulfide relative to the matrix. These spots were lined with dark ferrous phosphate, also present in accompanying submillimeter-scale “poppy seeds.” Both populations of spots may be explained by localized reductive dissolution and loss of ferric iron; at Cheyava Falls, reduction of iron and sulfate may have been coupled to the oxidation of organic matter, liberating phosphate (Hurowitz et al., 2025). The greenish spots in the Malgosa Crest abrasion patch in the Mount Spoonhead member seem especially similar in appearance to Earth’s reduction spots (Broz et al., 2025). Curiosity has also observed features resembling reduction spots, with white halos and dark cores, in the slightly pink Carolyn Shoemaker formation of Gale Crater (Seeger and Grotzinger, 2024).

On Earth, reduction spots are thought to form when microorganisms couple the oxidation of organic matter to the reduction of iron during sediment diagenesis (pre- or post-lithification), creating a redox front that expands radially from a microbial colony (e.g., Hofmann, 1990). However, an abiotic origin can rarely be excluded, and evidence for biogenicity is typically indirect. For example, Spinks et al. (2010) reported sulfide δ³⁴S values in a population of Triassic reduction spots markedly lighter than co-eval sulfates, suggestive of microbial sulfate reduction but within the range of (thermochemical) abiotic sulfate reduction. Similarly, in a global sample of Mesoproterozoic—Recent reduction spots, McMahon et al. (2018) found a pattern of progressive increase in δ²³⁸U from matrix to halo to core, consistent with the enzymatic fractionation of uranium isotopes by metal-reducing bacteria (Stylo et al., 2015) but not unequivocally so (Brown et al., 2018).

Because the redox reactions that occurred at Cheyava Falls could have been driven either by microorganisms or by abiotic chemistry, Hurowitz et al. (2025) describe them as “potential biosignatures.” A phenomenon is a “potential biosignature” if, first, “biological processes are a known possible explanation” and, second, “potential abiotic causes have not yet been reasonably explored and ruled out” (Gillen et al., 2023). Reduction spots meet the second criterion: Potential abiotic causes have not yet been reasonably explored, although experiments have shown that abiotic reductants can leach iron from sediment (Moulton, 1926; Hofmann, 1991). Arguably, however, reduction spots have not hitherto met the first criterion: It has never been demonstrated that biological processes can produce them.

This study, therefore, aimed to determine whether microbes can in fact produce reduction spots. We used Winogradsky columns to enrich natural mixed communities from pondwater and soil, expecting these communities to contain iron-reducing microorganisms, which are ubiquitous in near-surface environments on Earth (Weber et al., 2006). Samples from the Winogradsky columns were introduced to ferric iron-supplemented sediment slurries in an anaerobic chamber, mimicking the diagenetic conditions under which reduction spots are thought to form. We present evidence that these experiments successfully produced biogenic reduction spots.

2. Methods

2.2. Sediment bleaching experiments

In Experiment 1, 57 days after the columns were established, 1 mL of inoculum was transferred anaerobically by sterile Pasteur pipette from the upper region of each black layer in each column into an autoclaved 1-mm ball of absorbent cotton wool (CW) (see Fig. 1 and below). One cotton wool ball received 1 mL of sterile pond water instead of a microbial inoculum. In Experiment 2 (day 114), the black layer in columns SFL and PNFL was sampled as before. In addition, a sterile control was prepared whereby the inoculum from the column was passed through a 0.22 µm nylon syringe filter before soaking into the cotton ball. In Experiment 3 (day 137), the black layers in columns SFL and PNFL were again sampled as before. In Experiment 4, the cotton ball was removed from the SFL dish in Experiment 1 (80 days after the original setup) and transferred to a new dish. Finally, an experiment was performed using a pure culture of the iron-reducing bacterium Geobacter bemidjiensis cultivated on Geobacter Medium 579.

OPEN IN VIEWER

FIG. 1.

Schematic illustration of experimental set-up. In an anaerobic chamber, material from the reducing zone of Winogradsky columns (initially prepared with pondwater and soil) was introduced to balls of cotton wool, which were placed on a slurry of ferric-iron-rich sediment and sterile pondwater in Petri dishes in an anaerobic chamber under 20% CO₂, 80% N₂. The sediment mix consisted of 98 wt% quartz powder and 2 wt% synthetic ferrihydrite. Dishes were sealed with parafilm. See Supplementary Figure S1 for photographs of the Winogradsky columns.

Under a laminar flow hood, “quartz sand” powder (particle size <63 µm, Sigma-Aldrich; sterilized at 150°C for 16 h in a Teflon bottle) was mixed with 2 wt% ferrihydrite following Vixseboxse et al. (2024) and dispensed into 90-mm-diameter Petri dishes (22 g per dish; Experiments 1 and 2) or 40-mm dishes (3 g per dish; Experiment 3). Each dish was moistened with 10 mL (3mL in smaller dishes) of sterilized (filtered <0.22 µm) diluted pond water (collected from Blackford Pond, 11 June 2025) using a sterilized spray bottle, producing a < 1 mm supernatant above the saturated sediment. Dishes exchanged gas for 3 days in an anaerobic chamber (Coy; Vinyl Anaerobic Airlock Chamber, MI, USA; 20% CO₂ and 80% N₂ atmosphere).

In the anaerobic chamber, the inoculated cotton balls were placed gently on top of the sediment in the middle of each dish using sterile tweezers. All dishes were parafilm-sealed. Water lost by evaporation was replenished twice monthly to maintain the supernatant. Dishes were infrequently removed from the anaerobic chamber during the experiment for photography (∼3 min) and flatbed scanning (∼5 min, using a XEROX AltaLink C8035 scanner).

The methods used for DNA extraction and sequencing, resin impregnation, and electron microscopy (including energy-dispersive X-ray spectroscopy, EDX) are described in the Supplementary Data.

Our observations confirm that bleaching occurred via the reductive dissolution and diffusive loss of ferrihydrite. Orange-brown blemishes near the bleached zones and orange-brown crusts appearing on dish upper surfaces (Supplementary Fig. S2) probably indicate re-oxidation of iron derived from the ferrihydrite. Vixseboxse et al. (2024) produced comparable orange halos around decaying organic matter in similar experiments. No other mineralization was observed.

We are not aware of any previous report of the production of sedimentary reduction spots in the laboratory. Moulton (1926) found that crude oil was unable to bleach powdered ferruginous sediment. Hofmann (1991) successfully leached hematite from sedimentary rock with dithionite, a powerful artificial reducing agent. More generally, it is well known that sulfides and organic acids can abiotically reduce and dissolve Fe(III), especially in poorly or noncrystalline phases, in days (Moulton, 1926; Jauregui and Reisenauer, 1982; Miller et al., 1986; Torres et al., 1990; Ionescu et al., 2015), (including ferrihydrite; Miller et al., 1986). Organic acids adsorbed onto Fe-oxide surfaces form inner-sphere metal chelates; reduction via ligand-to-metal electron transfer results in the dissolution of the iron oxide and removal of ligand-bound Fe(II) complexes from the Fe-oxide surface, driving continued dissolution.

Other than the activity of microbes, our initial experiments contained at least two sources of sulfides and organic acids, namely, the 1 mL of material transferred from the Winogradsky columns and the pondwater in which all experiments were performed. Nevertheless, we infer that bleaching observed in our experiments was microbially mediated, for six reasons:

1.

Of the four dishes inoculated from the same layer in column SFL, spots appeared in three and failed to appear only in the one dish whose inoculum was filter-sterilized, a process that would have removed cells but not organic acids or other reductants.

2.

The sequencing data revealed abundant bacteria of iron-reducing genera in the sediment above the bleached areas in the SFL dish from Experiment 1 (64 days after inoculation); such genera were especially dominant in the peripheral area.

3.

All samples used to inoculate all six Petri dishes were reducing, sulfidic, and rich in organic matter, being taken from the black zone of Winogradsky columns. Nevertheless, inocula from only two columns caused bleaching, which must, therefore, require a more specific (and temperamental) agent than organic matter alone, for example, certain microbial strains.

4.

Bleaching apparently accelerated in Experiment 1, suggesting a lag period and/or gradual propagation of bleaching downward through the sediment, both typical of microbial growth. Abiotic solubilization of iron would have occurred rapidly after the introduction of material to the dishes and slowly thereafter (e.g., Miller et al., 1986). In contrast, dish PNFL bleached faster between days 22 and 36 than between days 0 and 22 (Fig. 1). Visible bleaching of the margin of dish SFL did not begin until after day 22. Similar delays occurred in the other experiments.

5.

The late marginal bleaching observed in Experiment 1 dish SFL occurred at multiple discrete millimeter-scale points as if induced by immobile localized agents such as microbial colonies (Fig. 3a). Abiotic organic acids would not have been spatially focused in this way.

6.

When the cotton ball was removed from Experiment 1 dish SFL and transferred to a new dish, a new bleached spot formed after several months, with no diminution of effect (Fig. 4). This suggests a successful “subculturing” of the original spot-forming population.

Although limited in size and scope, this study represents a first step towards a more exhaustive exploration of the mechanisms that might explain the occurrence of reduction spots in red rocks on Earth and Mars. Our general experimental approach is facile and easily replicated and extended. To enhance the realism of our experiments, the method could be improved by replacing granular ferrihydrite with a more cement-like phase precipitated directly onto the sand grains. It would also be worthwhile to incorporate pressure–temperature conditions, fluid compositions, and iron (oxyhydr)oxide phases characteristic of (later) diagenesis. Finally, future work should seek diagnostically useful differences between bacteriogenic and abiotic reduction spots and their associated morphological, geochemical, and mineralogical fingerprints.

5. Conclusion

Our results show that iron-reducing microorganisms can produce discrete, rounded, millimeter–centimeter-scale bleached spots in ferruginous sediments, suggesting that those found commonly in red beds on Earth—and those recently observed by Perseverance on Mars—are indeed “potential biosignatures.” The demonstration that microbes can produce reduction spots under some conditions does not imply that all reduction spots are biogenic. It is now essential to investigate potential nonbiological pathways to reduction spot formation on Earth and Mars. Nevertheless, our results lend weight to the suggestion that reduction spots in Neretva Vallis may be the first visible sign of life discovered on another world.