4D imaging reveals mechanisms of clay-carbon protection and release

Soils are a major carbon reservoir exchanging ~60 Gt C annually with the atmosphere and absorbing about 20% of anthropogenic emissions, so changes in soil carbon storage strongly affect the global carbon cycle and climate. Clay minerals, especially smectites and nano-crystalline Fe/Al oxides, are key regulators of soil carbon storage and turnover, yet the mechanisms of mineral-associated protection in the presence of microbes and extracellular enzymes are unclear. A prevailing hypothesis is that sorption of organic molecules to clay surfaces protects carbon from microbial decomposition, and many soil carbon models thus include a clay-controlled protected pool represented as reversible sorption. However, observations of rapid release of clay-associated carbon (priming) following inputs of low molecular-weight sugars suggest that microbial processes and extracellular enzymes may directly undermine mineral protection, a coupling not captured in many models. Direct, real-time visualization of carbon dynamics within clay micro-aggregates has been lacking due to opacity and reliance on destructive, static imaging. This study aims to directly image and quantify the dynamics of sorption and release of organic carbon within smectite clay micro-aggregates in the presence of bacteria and exoenzymes, to determine when and how clay provides protection and how enzymes can trigger carbon release, with implications for improving soil carbon models.

Prior work recognizes mineral protection as a key control on soil organic carbon, with correlations between clay content (notably smectites) and carbon storage. Conceptual models often posit that smaller organic molecules have greater access to mineral surfaces and thus greater mineral protection, and many process-based soil carbon models treat mineral sorption/desorption as reversible and independent from microbial activity. Conversely, priming studies demonstrate that additions of low molecular-weight substrates can stimulate microbial activity and lead to rapid loss of mineral-associated carbon over days, implicating microbial and extracellular enzymatic processes. Observational constraints inside opaque clay aggregates have relied on static destructive imaging (2D/3D snapshots), limiting mechanistic insight into sorption/desorption dynamics. The authors position their work to bridge these gaps by real-time 4D imaging to test how molecular weight affects sorption reversibility, how bacteria interact spatially with clay aggregates, and whether exoenzymes can penetrate aggregates to mobilize protected carbon.

The study employed a soil-on-a-chip microfluidic platform with a transparent synthetic smectite (laponite) to directly image carbon sorption and release in 4D (3D space + time) using confocal microscopy. Clay aggregate preparation: Laponite-RD powder (25 nm diameter, 1 nm thick disks; average pore size ~2 nm) was mixed in M9-based buffer (pH ~7) with micronutrients using a vortex mixer (2000 rpm, ~1 min), then equilibrated 24 h at room temperature to form micron-scale aggregates (10–100 µm). Microfluidic devices (PDMS bonded to #1.5 coverslip) were fabricated via soft lithography; channels were 40 µm high, 300–400 µm wide, ~5 mm long. Clay suspensions were manually injected at several mL/min to load and immobilize aggregates; subsequent flows were driven by syringe pump at 1 mL/h. Sorption/desorption imaging: Two fluorescent organics were used: FITC-labeled dextran (3–5 kDa; also 20 kDa and 70 kDa in some experiments) and fluorescent glucose (2-NBDG, 340 Da). After establishing aggregates, a solution with 0.05 g/L fluorescent organic was injected for hours to allow sorption to equilibration, followed by organic-free buffer for desorption. Confocal imaging (Leica TCS SP5) collected 3D stacks (horizontal resolution ~1.5 µm, vertical 1–2 µm) at 1-min intervals (sorption/desorption) or 10-min intervals (enzyme experiments). The average fluorescence within each aggregate was computed by delineating the aggregate boundary at saturation (intensity thresholding) and averaging pixel intensities; means and standard errors across multiple aggregates per channel were reported. Batch sorption isotherm: 10 mg laponite was mixed with 10 mL dextran solutions (3–5 kDa) spanning ~0.009–>2 g/L in glass vials, shaken (200 rpm, 3 days, room temperature, foil-wrapped). Supernatants were centrifuged (1350 × g, 10 min) and fluorescence quantified against calibration to obtain solution concentrations before (C0) and after equilibrium (Ce). Sorbed concentration on clay (Cs, g/g) was computed by mass balance. Bacteria-clay distribution: Pseudomonas aeruginosa PA14 expressing mCherry was cultured, then mixed with clay in M9 buffer containing D-glucose (4 g/L) and 2-NBDG (0.04 g/L), incubated 24 h at 37 °C in a culture well dish; confocal cross-sections visualized bacterial localization relative to clay aggregates. Enzyme release experiments: In a microfluidic “macropore” geometry, sequential injections were performed at hour-scale intervals: (1) 0.05 g/L green 3–5 kDa dextran, (2) 0.05 g/L red 20 kDa dextran, (3) 0.05 g/L green 70 kDa dextran, (4) no organics, (5) 2 g/L dextranase (Penicillium sp.; ~7 nm). Green/red fluorescence within immobile clay regions was tracked over time to assess uptake, reversibility, and enzyme-induced release. Controls and replicates verified reproducibility and accounted for photobleaching effects.

• Sorption dynamics by molecular weight: High molecular-weight sugars (FITC dextran 3–5 kDa) displayed quasi-irreversible sorption within micron-scale smectite aggregates: after switching to organic-free buffer, >50% of the initial average fluorescence remained in aggregates after ~50 h of desorption. In contrast, low molecular-weight sugar (fluorescent glucose, 340 Da) sorbed reversibly and desorbed rapidly upon removal from solution.
• Timescales: Dextran uptake reached apparent equilibrium within ~5 h; glucose sorption/desorption occurred on much shorter timescales (hours), fully reversible.
• Sorption isotherm: Batch experiments for 3–5 kDa dextran showed an S-shaped isotherm. A plateau sorbed loading Cs-plateau = 0.2 ± 0.1 g dextran per g clay was observed for Ce ≈ 0.009–2 g/L, consistent with a monolayer-equivalent coverage (estimated thickness h ≈ 2.1 ± 1.2 Å based on smectite specific surface area ~760 ± 40 m²/g and dextran density ~1.3 ± 0.3 g/cm³). For Ce > ~2 g/L, Cs increased sharply, consistent with secondary sorption via organic-rich phase formation (e.g., capillary condensation or surface-promoted aggregation). Low-MW glucose exhibited a linear isotherm without a plateau.
• Spatial microbial exclusion: Confocal imaging with Pseudomonas aeruginosa mCherry showed bacteria localized at aggregate peripheries, with negligible penetration into clay interiors, supporting physical protection of sorbed organic matter by size exclusion (bacteria micron-scale vs clay pores ~2 nm). This pattern was consistent across ≥4 replicates.
• Enzymatic release: Exoenzyme dextranase (~7 nm) penetrated clay aggregates and caused rapid (within several hours) decrease of both green (3–5 kDa) and red (20 kDa) fluorescence to near background, indicating enzymatic depolymerization of quasi-irreversibly sorbed dextran into smaller fragments that desorb readily. Injection sequence included 2 g/L dextranase after loading with 0.05 g/L dextrans (3–5, 20, 70 kDa); no significant desorption occurred during enzyme-free buffer flushing, confirming quasi-irreversibility without enzymes.
• Diffusion limitation with size: Larger dextran (20 kDa, 70 kDa) showed limited penetration with only aggregate edges labeled over experimental durations, indicating reduced diffusivity in clay; 70 kDa uptake was negligible within time frame.
• Conceptual and modeling implications: Findings reconcile mineral protection with priming by showing that (i) clay physically protects carbon via bacterial exclusion and strong sorption of high-MW compounds, and (ii) exoenzymes can breach this protection and mobilize carbon. The authors propose coupling exoenzymatic activity to release of clay-protected carbon in soil carbon models.

The study directly addresses how clay minerals protect organic carbon and under what conditions this protection can be compromised. High molecular-weight organics are quasi-irreversibly sorbed within smectite aggregates and spatially protected from bacteria that cannot access nanometer-scale pores, explaining mineral-associated carbon persistence. However, extracellular enzymes, unlike cells, can infiltrate aggregates, depolymerize sorbed macromolecules, and trigger rapid release of carbon into solution where microbes can consume it. This mechanistic picture links mineral protection and priming: inputs of labile substrates promote exoenzyme production, which in turn accelerates loss of clay-protected carbon, providing a positive feedback. The S-shaped isotherm with monolayer-equivalent plateau supports specific surface coverage at low Ce and suggests additional organic-rich phase formation at higher Ce. Overall, the findings indicate that mineral protection is dynamically regulated by microbial enzymatic activity and organic molecular size, implying that models must treat mineral-organic interactions and biotic processes as coupled rather than independent.

This work introduces 4D imaging of organic carbon dynamics within transparent smectite clay aggregates on a microfluidic chip, revealing that high-MW sugars are quasi-irreversibly sorbed and physically protected due to bacterial exclusion, yet remain vulnerable to exoenzymatic depolymerization that rapidly releases carbon. The sorption isotherm demonstrates a monolayer-equivalent plateau and secondary sorption at higher concentrations. These mechanistic insights reconcile observations of mineral protection and priming and motivate a revised soil carbon model structure that explicitly links exoenzymatic activity to the release of clay-associated carbon. Future research should extend soil-on-a-chip experiments to co-culture enzyme-producing microbes with clay in situ, explore the diversity and activity of exoenzymes and their interactions with various minerals and organic chemistries, and investigate the roles of water saturation and oxygen gradients on microbial respiration and carbon fate.

• Model system: Experiments used synthetic transparent smectite (laponite) and fluorescently labeled sugars, which simplify the complexity of natural soils and organic matter chemistries (e.g., diverse functional groups not explored).
• Organismal scope: Bacterial-clay spatial exclusion was demonstrated in culture dishes, not yet in fully integrated microfluidic co-cultures; only one bacterial species (P. aeruginosa) was tested, and enzyme was exogenously supplied (dextranase), not produced in situ by microbes.
• Temporal and environmental constraints: Observations span hours to days under water-saturated conditions with gas-permeable PDMS; effects of variable water content, redox/oxygen limitations, and temperature were not assessed.
• Analytical constraints: Confocal fluorescence is semi-quantitative and subject to photobleaching; fluorescence intensity is a proxy for concentration. CO2 production or direct mineral-bound carbon quantification was not measured.
• Molecular scope: Focus on sugars (glucose and dextran) and limited molecular weight range; other classes of organic matter and higher structural complexity may behave differently.
• Transport limitations: Diffusion dynamics inferred qualitatively; detailed pore-scale transport properties and enzyme kinetics within aggregates were not quantified.