Planting material
The five rice varieties used in the experiments included two allelopathic varieties (PI312777 and PI338046), two non-allelopathic varieties (PI502968 and IR55423), and variety IR80508-B-57-3-B which has not been tested for allelopathic effects. The allelopathic potential of these varieties is defined based on their ability to interfere with weed seed germination and growth (Khanh et al. 2007; Kong et al. 2008). Seeds of varieties PI312777 and PI338046 were obtained from Dale Bumpers National Rice Research Center, AR, USA. Seeds of varieties IR55423 (Apo) and IR80508-B-57-3-B (ApCr) were obtained from the International Rice Research Institute, Los Baños, The Philippines. Seeds of PI502968 (Rexmont) were obtained from the National Small Grain Collection, USDA-ARS, Idaho, USA. Rice seeds used in the experiments were surface-sterilized with 1% sodium hypochlorite (NaClO) for 30 min and soaked in sterilized distilled water for 24 h (Bi et al. 2007).
Extraction of rice-root derived compounds
Rice varieties PI312777, PI338046, Apo, and ApCr were tested for their BNI potential. Rice was grown in a capillary mat system to generate large quantities of living root tissues as described previously (Bertin et al. 2003; Czarnota 2001). Two watering regimes were used as treatments: (a) continuous watering (CW), and (b) water stressed (NW). Three replications per rice variety by watering regime treatment combination were used. The capillary mat systems seeded with surface sterilized seeds were placed in a controlled environment chamber with a temperature of 25 ± 1ºC and a photoperiod of 14/10 h light/dark. The capillary mat system was kept moistened with Milli-Q water until 14 days after germination (DAG) for the CW treatment; whereas, watering was discontinued at 9 DAG and allowed to gradually dry to impose the water stress treatment (NW) (Bertin et al. 2003). At 7 DAG, plants were watered once with half strength Hoagland’s nutrient solution (Hoagland and Arnon 1950). At 14 DAG, roots were harvested. The fresh roots were submerged in 25 ml of Milli-Q water for 30 min to extract water soluble root exudates (RE). The extracts were filtered using Whatman No 42 filter paper and stored at 4°C until needed. After recording fresh weights, roots were oven-dried at 50°C for 48 h. Dry weights of roots were recorded, roots were ground to a powder, and the water-soluble compounds were extracted with 30 ml Milli-Q water by placing containers on a shaker overnight at 120 rpm at room temperature followed by centrifuging and filtering the extracts through Whatman No 42 filter paper. This extract is referred to as root tissue (RT), hereafter. Solutions were stored at 4°C until needed to prepare stock solutions. Both RE and RT extracts were concentrated using a rotary evaporator, evaporated under vacuum to obtain the dry matter, and reconstituted in water to obtain 10 mg ml− 1 stock solutions, which were stored at -20°C until needed. Aseptic conditions were maintained throughout the RE and RT preparation process.
Osmotic potential of individual stock solutions of RE and RT was measured using a Fiske® 210 micro-sample osmometer (Advanced Instruments, Norwood, MA, USA). The pH of RE and RT at 0.50 and 1.0 mg ml− 1 dilution levels in water and P buffer medium (composition of the medium in g L− 1: 0.7 potassium dihydrogen phosphate (KH2PO4); 13.5 disodium phosphate (Na2HPO4); 0.5 sodium bicarbonate (NaHCO3); 0.1 magnesium sulfate (MgSO4.7H2O); 0.05 calcium chloride (CaCl2.2H2O); and 0.001 ferric EDTA (Fe-EDTA)) at pH 7.8 were measured using BDH® pH test strips (VWR, Arlington Heights, IL, USA).
The effect of rice-root derived compounds on nitrification
A pure culture of Nitrosomonas europaea (ATCC 19718) was used to assess the impact of different dilution levels of RE (0.05 and 0.50 mg ml− 1) and RT (0.05, 0.50, and 1.0 mg ml− 1) on the potential ammonia oxidation activity (PAOA). Activity of the N. europaea culture inoculated with sterile distilled water to replace the RE/RT volume served as the control. Nitrosomonas europaea cells were cultivated in NH4+ containing P-buffer medium at pH 7.8 (ammonium sulfate ((NH4)2SO4) added at 2.5 g L− 1) in an incubator shaker (250 rpm; temperature 30°C) (Subbarao et al. 2006b). After 14 days of incubation the N. europaea culture was used in the BNI bioassay. The population level of N. europaea on day 14 was 104 as determined by a most probable number technique (Rowe et al. 1977).
For the BNI bioassay, 400 µl of 14-day old N. europaea culture were mixed with 10, 100, or 200 µl of 10 mg ml− 1 stock solution of RE or RT and 1590, 1500, or 1400 µl of fresh NH4+ containing P-buffer medium to obtain 0.05, 0.50, and 1.0 mg ml− 1 concentrations of root-derived compounds, respectively. The mixture was incubated at 30°C while on a shaker at 250 rpm for 15 h. After 5, 10, and 15 h of incubation, 400 µl subsamples were collected and analyzed for nitrite colorimetrically using the sulfanilamide method (Shinn 1941). The BNI bioassay was performed in two laboratory replicates for each dilution level of the three stock solutions (RE/RT) from the combination of rice variety and water treatment. For the colorimetric analysis, a standard curve was prepared using sodium nitrite (Na2NO2). The BNI potential of rice was calculated as the percent reduction in PAOA by RE or RT at a given concentration compared with the PAOA of the control. Nitrapyrin, a commercially available nitrification inhibitor (N-serve® with 99% active ingredient) was used with the N. europaea culture in a test series (5, 10, 15, and 20 µg ml− 1) to compare to the BNI results (Zacherl and Amberger 1990).
Allelopathic potential of rice-root derived compounds
Due to its high sensitivity to allelochemicals, lettuce (Lactuca sativa L.) is commonly used in allelopathy bioassays (Elakovich and Wooten 1991). To assess the allelopathic potential of root-derived compounds, bioassays using lettuce (Harris® seeds, Rochester, NY, USA) were conducted. For each stock solution, a dilution series of RE (0.05 and 0.50 mg ml− 1) and RT (0.05, 0.50, and 1.0 mg ml− 1) were prepared in two laboratory replicates and used in bioassays. Ten lettuce seeds were placed on a Petri dish lined with a Whatman No.1 filter paper, which was moistened with 1.0 ml of aqueous extract at each respective dilution level. A control, which received 1.0 ml of Milli-Q water was included in the bioassay. Petri dishes were sealed and kept in a growth chamber with 12 h light/12 h dark cycle at 25°C for 4 d. Germination percentage and root and shoot lengths of lettuce seedlings were measured at the end of 4 days and percent inhibition was calculated based on seedling growth in the control treatment.
Root-derived substance extractions from rice growing on capillary mats, and subsequent bioassays with ammonia oxidizers and lettuce seeds were repeated once to confirm the repeatability of results.
Effect of biological interactions in the rhizosphere on the activity of nitrifiers
Microcosm construction
Microcosms were constructed using PVC couplings of 5.1 cm diameter and 3.8 cm height (Charlotte® pipe 1½ inch PVC coupling - Model #: PVC 00100 0800). A Nitex® 30 µm nylon mesh (Genesee Scientific, San Diego, CA, USA) was glued (vinyl fabric and plastic flexible adhesive) to the middle wall to divide the coupling into two compartments; a top rhizosphere compartment and a bottom soil compartment. The nylon membrane allowed free nutrient flow but restricted roots to the rhizosphere compartment. A soil sample collected at 0–10 cm depth from a rice paddy in the Cornell Botanic Gardens, Ithaca, NY, USA, was air-dried and sieved through a 2 mm mesh sieve. To increase the efficiency of separating rhizosphere soil from roots and to avoid crack formation and soil crusting during the experiment, soil was mixed with sterilized sand to achieve 70% sand, 12% clay, and 18% silt composition. The two compartments of the microcosm were filled with soil. The bottom compartment was sealed and a small hole was created and plugged with cotton to facilitate drainage. Each microcosm was placed in a water reservoir, which helped to regulate the soil moisture level in the rhizosphere compartment. Two soil moisture treatments were imposed, (i) continuously saturated soil (TC) and (ii) unsaturated soil kept at near field capacity (TU), starting from six days prior to seeding rice. Microcosms were incubated at 25ºC.
Four rice varieties were used in the microcosm experiment including ApCr, PI312777, and PI338046, and Apo was replaced with Rexmont, a well-known non-allelopathic rice variety, due to the availability of seeds. Rice seeds were surface-sterilized with 1% NaClO for 30 min, rinsed with sterile-distilled water and pre-germinated in petri-dishes lined with wet filter papers for three days. Six germinated seeds with equal radicle lengths were planted in the rhizosphere compartment. Three replicates per variety per moisture treatment combination were used. The microcosms were placed in a growth chamber and plants were subjected to 28°C day and 25°C night temperature and a photoperiod of 14/10 h light/dark. Plants were fertilized with half strength Hoagland’s nutrient solution at 7 and 10 DAG. Soil and plant sampling were performed at 14 DAG. Soil was separated from roots and placed in plastic sampling bags and immediately analyzed for potential nitrification rate (PNR) and plant available N content. A subsample of rhizosphere soil was stored at -20˚C to extract soil DNA for microbial community profiling. Shoots were separated from roots and oven-dried at 65°C for 48 h.
Soil and plant analyses
The potential nitrification assay was conducted using the shaken-slurry method as described by Hart et al. (1994) modified for sample size. In brief, for each sample, 3 g of soil (oven dry weight equivalent) was placed into a 125 ml Erlenmeyer flask and a 25 ml aliquot of phosphorus buffer solution containing 1.5 mM (NH4)2SO4 was added. Flasks were shaken at 180 rpm on an orbital shaker and aliquots were removed at 1.5, 3, and 5 h after incubation, and analyzed colorimetrically for NO3−-N as described by Cataldo et al. (1975).
Available inorganic nitrogen in soil from the microcosm experiment was extracted with 2 M potassium chloride (KCl) and analyzed colorimetrically for NH4+ and NO3− using an auto-analyzer (Seal Analytical Inc., Mequon, WI, USA). Shoots were oven-dried, weighed, and analyzed for total C and N using the dry combustion method (automatic carbon-nitrogen analyzer NC2100, EA/NA 1110, ThermoQuest Italia S.p.A., Milan, Italy). The physiological NUE (PNUE) of the plants was calculated as dry biomass accumulated per gram of nitrogen acquired by the plant.
Terminal restriction fragment length polymorphism (T-RFLP) analysis
DNA was extracted from sampled soils using the PowerSoil™ DNA extraction kit (MoBio Laboratories, Carlsbad, CA) and used for PCR to perform T-RFLP analysis. Bacterial 16S rRNA genes were amplified with universal primers 27F and 1492R (Culman et al. 2008). Archaeal 16S rRNA genes were amplified with the universal primers 109F and 912R (Ramakrishnan et al. 2000). Ammonia-oxidizing bacteria amoA genes were amplified using primers amoA-1F and amoA-2R (Rotthauwe et al. 1997) and archaeal amoA were amplified using primers ArchamoAF and ArchamoAR (Chen et al. 2008). All the primer sequences and amplifying conditions are provided in Supplementary Table 1 (S1).
One primer from each pair was end-labeled with a 6-FAM fluorophore to enable use of the PCR products to generate T-RFLP ‘fingerprint’ profiles (S1). Accordingly, 27F, 109F, amoA-2R and ArchamoAF were fluorescently labeled. Reactions were performed in a 50 µl reaction volume with approximately 100 ng DNA per reaction. For each soil DNA extract, a PCR reaction was performed in triplicate and products were pooled, vacuum-dried, and reconstituted in sterile molecular grade water to obtain 20 ng DNA µl− 1. PCR products were restricted using the HhaI enzyme for total bacteria and archaea (16S rRNA gene amplicons) and AOA. Enzyme TaqI was used to digest AOB amplicons. Restriction digest products were column-purified using an EdgeBio purification plate (Applied Biosystems, Foster City, CA) and were lyophilized for a final time. DNA was re-suspended in a 10 µl mix containing 9.85 µl of formamide and 0.15 µl of Liz 500 size standard (Applied Biosystems) and terminal restriction fragments (T-RFs) size analysis was performed using an ABI 3730 electrophoretic capillary sequencer (Applied Biosystems) located at the Cornell Biotechnology Core Laboratory.
Statistical analysis
Data generated from the allelopathy, BNI, and microcosm experiments were tested for normality and used in analysis of variance (ANOVA) with the generalized linear model (GLM) procedure in a factorial design to determine significance of treatment effects using JMP 8.0® software (SAS Institute Inc., Cary, NC, USA). Rice variety, moisture treatment, and dilution level of root-derived compounds were used as the grouping factors in the allelopathy and BNI studies. Rice variety and moisture treatment were used as the grouping factors in the microcosm study. When treatment effects were found to be significant at p < 0.05, means were compared using the LSD mean separation technique (p < 0.05).
T-RFLP profiles were analyzed using GeneMapper Software v 3.0 (Applied Biosystems). T-RFLP profiles were further analyzed using T-REX, a web-based tool (http://trex.biohpc.org/), as described previously (Culman et al. 2009). To align T-RFLP profiles, we used a clustering threshold of five for amoA and two for the 16S rRNA gene amplicons, based on positive controls used for correction of shifts in peaks due to limitations in the T-RFLP method related to the fluorescent signals (Culman et al. 2009). Compositional differences were investigated using the Additive Main Effects and Multiplicative Interaction Model (AMMI) via MatModel 3.0 software (Culman et al. 2008). Interactive principal component plots were developed after considering the values in the ANOVA table given in the AMMI output. Data sets that had high error sum of squares values in the ANOVA were omitted in the data interpretation (Culman et al. 2008).