UC Berkeley

Organic matter decomposition is a central step in the global carbon cycle. Tropical forests generally have high net primary productivity (NPP) and a large percentage of the global carbon stock, making them an important component of the global carbon budget. Organic matter decomposition ultimately produces carbon dioxide (CO₂) and methane (CH₄), both important greenhouse gases. Rates of organic carbon mineralization are affected by soil physical parameters, such as fluctuating redox conditions, and biological parameters, such as microbial community composition and the microbial genetic potential for organic carbon mineralization. Tropical forest ecosystems are predicted to become warmer and drier due to global climate change; such changes would affect CO₂ and CH₄ efflux from tropical forest soils. Thus, understanding how physical and biological factors sensitive to climate affect organic matter mineralization to CO_{2 and CH4 in these soils is important.}

The possible microbial controllers of CH₄ emissions in upland tropical forest soils were evaluated in Chapter 1. Recent evidence has suggested that some upland tropical forest soils may be CH₄ sources. This has important implications for climate models because most models consider tropical forest soils to be CH₄ sinks. Here, two microcosm-based experiments were used to identify possible physical (edaphic) and biological (microbial community size and diversity) factors that could influence CH₄ emissions in three upland tropical forest soils from the Luquillo Experimental Forest in Puerto Rico. Edaphic factors, quantification of functional gene markers by Polymerase Chain Reaction (PCR - mcrA for methanogens, pmoA for CH₄ oxidizers), and community composition (clone libraries of mcrA and pmoA) were correlated to cumulative CH₄ production and oxidation to determine the relative importance of each to CH₄ emissions in these soils. Unsurprisingly, edaphic factors including soil moisture and mean annual precipitation drive CH₄ production in these soils. Results indicate that in soils with high soil moisture, methanogens are more abundant, leading to greater potential CH₄ production. Methanogen community diversity did not significantly affect CH₄ production. However, mcrA sequences representing the methanogen genera, Methanosaeta and Methanosphaera were associated with high CH₄ production. Potential CH₄ oxidation was unaffected by any edaphic factor tested. However, potential CH₄ oxidation was correlated to the estimated size of the CH₄ oxidizing community (as measured by qPCR of the pmoA gene). In particular, greater potential CH₄ oxidation was observed in soils with greater Type I pmoA gene abundance. Rates of CH₄ oxidation were greatest in soils with the highest rates of CH₄ production. The results presented indicate that conditions conductive to methanogen activity (saturated soils) are key to the CH₄ production rates in these soils and that specific genera of methanogens are associated with the highest rates of production. The size of the CH₄ oxidizing community is associated with greater CH₄ oxidation in these soils and overall, CH₄ oxidation potentials were primarily dependent on the rates of CH₄ production.

The focus of Chapters 2, 3 and 4 shifts to the effects of redox conditions (oxygen availability), community profiles of microbial genes involved in decomposition processes (carbon degradation potential), and bacterial community composition on mineralization of organic matter to CO₂ and CH₄. High rates of plant material decomposition have been measured in the soil of the Luquillo Experimental Forest in Puerto Rico. These high rates have been attributed to the consistently warm and moist conditions that these soils experience. These soils also experience redox fluctuations driven in part by frequent rain events and high biological activity. In Chapters 2 and 3, I investigated the following questions: what is the effect of redox condition on organic carbon mineralization and carbon degradation potential, and are there links between mineralization of organic matter, carbon degradation potential, and bacterial community composition. I hypothesized that under fluctuating redox conditions, both aerobic and anaerobic decomposition processes occur and are synergistic, leading to greater rates of organic carbon mineralization than under static oxic or anoxic conditions. To address these questions and hypothesis, I added ¹³C-labeled organic carbon (either plant material, Avena barbata, or cellulose) as a substrate to soils in two microcosm-based experiments. In these experiments, the soil microcosms were either amended with the ¹³C-labeled carbon or left unamended and incubated under three redox conditions; oxic, anoxic, or alternating four-day oxic-anoxic fluctuation. The timescale of the two experiments was different; the A. barbata litter-amended experiment was 38 days, and the cellulose-amended experiment was 22 days. Headspace gases analyzed over time were used to measure cumulative CO₂ and CH₄ and the production of ¹³C-labeled CO₂ and CH₄. Soil microcosms were destructively sampled over time to allow determination of microbial community abundance by qPCR. Final time point samples were used to analyze the effect of redox condition on carbon and cellulose degradation potentials (by GeoChip 4) and bacterial community composition (by 454 Pyrosequencing). Degradation potential refers to the composition and abundance of gene probes on the GeoChip 4 microarray; for carbon degradation potential, the carbon degradation and organic remediation gene probes were analyzed and for cellulose degradation potential, the gene probes involved in cellulose degradation were analyzed. In Chapter 4, the results from the A. barbata litter and the cellulose addition experiments (Chapters 2 and 3) are compared to determine the importance of redox condition, degradation potential and bacterial community composition as controllers of mineralization of a complex plant litter compared to a purified component of plant litter, cellulose.

The influence of redox conditions on the mineralization of plant litter, microbial community abundance, bacterial composition, and carbon degradation potential in a wet, tropical forest soil is presented in Chapter 2. Redox conditions strongly influenced plant litter mineralization to CO₂ and CH₄. Mineralization of ¹³C-labeled A. barbata litter to CO₂ was greatest under fluctuating redox conditions. However, because substantial CH₄ was produced over 38 days under the anoxic condition, the mineralization of A. barbata litter to ¹³C-gas was not significantly different under the anoxic and the fluctuating redox conditions. Carbon degradation potential was also influenced by redox condition; the carbon degradation potential from samples under anoxic conditions was distinct from the carbon degradation potential from samples under oxic conditions. However, carbon degradation potential was not correlated to mineralization. Redox conditions did not affect bacterial community composition or microbial community abundances. However, there were substantially more Firmicutes sequences detected under anoxic conditions in the presence of added A. barbata litter. In addition, bacterial community composition was correlated to indices of mineralization (cumulative CO₂, cumulative CH₄ and ¹³C-CO₂ production). The release of soluble sugars from A. barbata litter probably enabled the activity of bacteria capable of rapid growth on easily degradable carbon sources, such as Firmicutes, thus affecting the bacterial community composition. Bacterial community composition was not correlated to carbon degradation potential, suggesting that the bacterial community composition is functionally redundant. There appears to be distinguishable aerobic and an anaerobic decomposition potentials which is consistent with a synergy between them under fluctuating redox conditions. Indeed, the ¹³C-CO₂ produced from the added plant litter was significantly greater under the fluctuating redox condition.

In Chapter 3, the influence of redox condition on the mineralization of cellulose, microbial abundance, bacterial community composition, and cellulose degradation potential is assessed. Cellulose is one of the most abundant biopolymers on earth. It is also considered a relatively recalcitrant component of plant material. There are two main microbial metabolic strategies to degrade cellulose: cell-free extracellular enzymes and cellulosomes, cell-associated supra-molecular structures. Cellulosomes are produced under anoxic conditions by both bacteria and fungi whereas cell-free extracellular cellulose degrading enzymes are produced under both oxic and anoxic conditions by bacteria and fungi. Redox condition influenced cellulose mineralization; cellulose mineralization was significantly greater under the static anoxic condition. This result indicates that anaerobic microbes play an important role in cellulose degradation in this soil. Redox condition also affected the cellulose degradation potential when the static oxic samples were compared to the static anoxic samples. However, when the fluctuating redox samples were included in the statistical analysis, redox condition did not affect the total detected cellulose degradation potential. The majority of the detected cellulose degradation potential gene probes were glycosyl hydrolases which are responsible for degrading cellulose fibers to monomer and dimer sugars. Glycosyl hydrolases are grouped into large families based on sequence homology and not microbial origin or whether they are produced under oxic or anoxic conditions. Also glycosyl hydolases do not require molecular O2 for function. Therefore, it is not surprising that overall redox condition did not affect the cellulose degradation potential. The organization of glycosyl hydrolase families also explains why carbon degradation potential and bacterial community structure were not correlated. Neither carbon degradation potential nor bacterial community composition were correlated to cellulose mineralization. However, bacterial community composition was correlated to cumulative CO₂ which suggests that in this experiment, bacterial community composition was important to mineralization of other allochthonous carbon sources. Cellulose mineralization and cellulose degradation potential in this soil appears to be largely uncoupled from bacterial community composition. Cellulose degradation in these soils may involve synergy between glycosyl hydrolases, but potential synergy in aerobic and anaerobic cellulose degradation was not supported.

In Chapter 4, the results from plant litter mineralization (Chapter 2) and cellulose mineralization (Chapter 3) are compared to assess differences and similarities in the mineralization and degradation potential of these organic carbon sources. Gas measurements from the A. barbata litter amended experiment were recalculated using data taken up to day 22 so that they could be compared to the gas measurements of the 22 day cellulose-amended experiment. GeoChip data from each experiment was re-analyzed so that a carbon degradation and a cellulose degradation potential from each experiment could be compared. Likewise, the bacterial community composition data from the litter-amended experimental samples were rarified to the extent of the samples from the cellulose-amended experiment so that they could be compared. In both experiments, redox conditions affected the mineralization to CO₂ and CH₄. In particular, under the anoxic condition, mineralization of cellulose to CO₂ was greater than under the other conditions tested, and mineralization of A. barbata litter to CO₂ was substantial. Only slight mineralization of A. barbata litter to CH₄ was observed after 22 days, but the CH₄ observed was greater under the anoxic conditions. CH₄ production was below the limit of detection in the cellulose-amended experiment. The substantial mineralization of both A. barbata litter and cellulose to CO₂ suggests that anaerobic microorganisms are crucial in decomposing plant litter and cellulose in this soil. Redox condition also affected the carbon and cellulose degradation potentials found in both experiments. A distinguishable aerobic and anaerobic carbon degradation potential was observed in the A. barbata litter- and the cellulose-amended experiments, and a distinguishable aerobic and anaerobic cellulose degradation potential was observed only in the A. barbata litter-amended experiment. A distinguishable aerobic and an anaerobic cellulose degradation potential was not observed in the cellulose-amended experiment. The observation of distinguishable aerobic and an aerobic carbon degradation potentials indicates that under fluctuating redox conditions, these degradation potentials could work synergistically.