Bacterial succession during 500 years of soil development under agricultural use

Abstract

Soil bacterial succession under intensive anthropogenic disturbances is not well known. Using terminal restriction fragment length polymorphisms and 454 pyrosequencing of 16S rRNA genes, this study investigated how soil bacterial diversity and community structure changed under two agricultural land uses (paddy rice and upland cropping) in relation to soil development along a 500-year chronosequence created by intermittent reclamation of estuarine salt marshes. Multivariate analysis revealed orderly changes in soil physicochemical properties and bacterial community structure with time, confirming the occurrence of soil development and bacterial succession. Patterns of soil development and bacterial succession resembled each other, with recent land uses affecting their trajectories but not the overall direction. Succession of bacterial community structure was mainly associated with changes in α-Proteobacteria and Verrucomicrobia. Two stages of bacterial succession were observed, a dramatic-succession stage during the first several decades when bacterial diversity increased evidently and bacterial community structure changed rapidly, and a long gradual-succession stage that lasted for centuries. Canonical correspondence analysis identified soil Na⁺, potentially mineralizable nitrogen, total phosphorous, and crystallinity of iron oxyhydrates as potential environmental drivers of bacterial succession. To conclude, orderly succession of soil bacterial communities occurred along with the long-term development of agroecosystems, which in turn was associated with soil physicochemical changes over time.

Appendix: Methods of soil physicochemical analyses

Soil bulk density (BD) was determined by oven-drying soil cores of a fixed volume at 105 °C to constant weight. Soil particle size distribution was analyzed with a LS 230 laser particle size analyzer (Beckman, USA) and the mean particle diameter (MPD) was calculated using the software equipped with LS 230. Aggregate distribution analysis was performed by wet-sieving through a series of sieves with a vibratory sieve-shaker (Analysette 3, Fritsch, Germany), and the obtained data were used to calculate mean weight diameter (MWD). Soil pH was measured on soil slurry at 2.5:1 water: soil ratio using a glass electrode. Carbonate content (IC) was determined by back-titrating soils neutralized with excessive 1 M HCl. Soil salinity was measured with a platinum electrode in the supernatant of soil slurries at 5:1 water: soil ratio and expressed as the percentage of total water-soluble salts on the dry weight base. Soluble cations (Na⁺, K⁺, Ca²⁺ and Mg²⁺) were extracted with 1 M NH₄OAc at pH 7.0. Free Fe and Al oxyhydrates (Fe_d and Al_d, respectively) were extracted with citrate–dithionite–bicarbonate (DCB), and amorphous Fe and Al oxyhydrates (Fe_o and Al_o) with oxalic acid–ammonium oxalate. Complexed Fe and Al (Fe_p and Al_p) were extracted with sodium pyrophosphate at pH 8.5. The extracted Na, K, Ca, Mg, Fe and Al were then measured with a P-4010 inductively coupled plasma (ICP) spectrometer (Hitachi Ltd., Japan). Total P (TP) was also measured with ICP after fusion with lithium metaborate at 1,000 °C. Available P (P_o) was extracted with 0.5 M NaHCO₃ at pH 8.5 and measured by colorimetry. Cation exchange capacity (CEC) was measured with the ammonium acetate method. Soil organic carbon (SOC) was measured with a TOC analyzer (Analytikjena HT1300, Germany) after removing soil carbonates with 1 M HCl. Labile carbon (LC) was estimated as the SOC fraction oxidizable by 333 mmol KMnO₄. Potentially mineralizable carbon (PMC) was determined by a 28-day incubation of 45 g field moist soils at 25 °C. The soil was placed in a 50-ml beaker, together with a plastic vial containing 20 ml of 1 M NaOH. The amount of CO₂–C trapped in NaOH (i.e., PMC) was quantified by titration with 0.5 M HCl after precipitation of Na₂CO₃ with BaCl₂. Potential mineralizable nitrogen (PMN) was determined as changes in the sum of NH₄ ⁺ and NO₃ ⁻ after incubation at 25 °C for 28 days. Microbial biomass carbon (MBC) was determined by the CHCl₃ fumigation-extraction method. Total nitrogen (TN) was determined with a C/N elemental analyzer (FlashEA 1112 NC analyzer, Thermo, Italy). Inorganic nitrogen (NH₄ ⁺, NO₃ ⁻) was extracted by 2 M KCl and measured with a discrete auto analyzer (Smartchem 200, Westco, France). Magnetic susceptibility (MS) was determined with a magnetization meter (MS-2B, Bartington, UK) at the frequency of 0.47 kHz.