Long-term use of green manure legume and chemical fertiliser affect soil bacterial community structures but not the rate of soil nitrate decrease when excess carbon and nitrogen are applied.
Nitrogen (N) is one of the most important nutrients for plant growth in soils. To maintain agricultural productivity in soils, chemical N fertiliser is commonly used. Alternatively, green manure legumes have been receiving attention as an N source to sustain plant productivity in soils because they establish a symbiotic relationship with N-fixing microbes and increase the amount of soil-N when they decompose. Green manure legumes are often planted immediately before, or together with, edible crops (de Vries and Bardgett 2012).
N is also an important nutrient for soil microbes to maintain their activities. Microbes can be sinks for N and the N within live microbes (biomass N) cannot be used by plants. The incorporation of plant-available N into microbes is known as N immobilisation and high N immobilisation activity often results in the loss of agricultural productivity due to reduced plant-available N. N immobilisation by microbes is dependent on many factors, such as the soil carbon (C) to N ratio and microbial activity (Barrett and Burke 2000; de Vries and Bardgett 2012). Generally, with increasing soil organic matter content, the potential of N immobilisation increases, and the quality of the C source in the soil is also an important factor controlling N immobilisation rates (Chen et al. 2014). For example, residues with C : N ratios between 9 and 23 resulted in net mineralisation (organic N decomposed to inorganic N), whereas residues with higher C ; N ratios (between 47 and 99) resulted in net immobilisation (Chen et al. 2014).
Thus, the use of green manure legumes may increase the amount of available N in soils, but it may also affect soil microbial N immobilisation potential. A previous study showed that the application of plant residue (wheat straw) reduced the amount of mineralised N compared with non-residue treatment due to N immobilisation (Mary et al. 1996). Other studies showed that vegetation consisting of high N content promoted N mineralisation more than N immobilisation (e.g. Orwin et al. 2010). Previous studies of mineralisation and immobilisation have not monitored the responsible microbial community, and it is still uncertain whether the long-term use of legumes alters soil microbial N immobilisation potentials.
A better understanding of the changes in soil microbial community structures may be required to better manage the long-term use of legumes and the depletion of available N in soils when high C : N ratio plant residues are added to the soils. It is known that soil microbes responsible for N cycles are strongly affected by the plant species (Mao et al. 2011). Blasko et al. (2013) also reported that fertilisation with ammonium nitrate increased Gram-positive bacteria more than Gram-negative bacteria and that N application decreased the abundance of micro-organisms. Changes in microbial structure, possibly caused using legume species, may affect plant-soil competition for available N.
Studies focusing on changes in soil microbial structures with different management histories (legume and chemical fertiliser) and the magnitude of the depletion of available N due to the addition of high C : N ratio plant residues are needed to efficiently use the green manure legume-derived N for subsequent crops. Thus, the aims of the present study were to (1) evaluate the effects of long-term legume application on microbial community structures and diversity and (2) compare the rate of nitrate decrease in soils with a history of legume or chemical fertiliser when C and N are added to the soils.
Materials and methods
Soil sampling sites
Soil (5 cm depth) was collected from a plot trial in an experimental greenhouse at the Field Science Center for Northern Biosphere, Hokkaido University, Sapporo, Japan (43[degrees]3'N, 141[degrees]20'E), in October 2014, after the harvest of the tomato crop. Soils were sampled both as intact cores (100 [cm.sup.3]) and in plastic bags. More details of this tomato trial have been provided elsewhere (Araki et al. 2009). Briefly, the plot trial consisted of five treatments, with three replicates: (1) chemical fertiliser (CF); (2) hairy vetch (HV; Vicia villosa) mulch; (3) oats (Avena strigosa) mulch; (4) HV plus oats mulch; and (5) hairy vetch incorporated. The plot trial had been established in 2006, and every year green manure (HV and/or oats) was planted during early spring (April), with tomatoes planted in June following surface cutting and mulching (or incorporation) of the green manure. In the present study, soils were sampled from only the CF and HV mulch plots (two treatments x three replicates) for laboratory-based incubation experiments, as detailed below.
From the establishment of this greenhouse trial until the soil sampling for the present study in 2014, the CF and HV plots received 240 and 120 kg N [ha.sup.-1] [year.sup.-1] respectively as N chemical fertiliser (20% and 80% of total N fertiliser was applied by a fast-release fertiliser (ammonium sulfate) and a slow-release fertiliser (polymer-coated urea; LPCOTES100; JCAM AGRI, Tokyo, Japan) respectively) in July, at the time tomatoes were planted, in 2006 and 2007. From 2008 to 2014, the same amount of N was applied on the soil surface in the CF plots, but the amount was reduced to 80 kg N [ha.sup.-1] [year.sup.-1] in the HV plots. The dry weight of the applied mulched HV was 4.7 [+ or -] 1.61 [ha.sup.-1] [year.sup.-1] from 2010 to 2012. This was equivalent to 203 kg N [ha.sup.-1] [year.sup.-1]. Phosphorus (P) and potassium (K) were also added to the soil surface at the time tomatoes were planted (200 kg P [ha.sup.-1] [year.sup.-1] as fused magnesium phosphate and 200 kg K [ha.sup.-1] [year.sup.-1] as potassium sulfate). The yield of tomatoes was generally greater for the CF than HV plots (47 vs 41 t [ha.sup.-1] [year.sup.-1] respectively, on average, from 2009 to 2012). Further information on the trial is available elsewhere (Araki et al. 2009).
Soil physical and chemical properties
The properties of the soils used in the present study are given in Table 1. For soil samples from each plot, bulk density and the water-filled pore space (WFPS) were determined by oven drying intact soil cores for >48 h. The particle density of the soil was assumed to be 2.65 g [cm.sup.-3]. Total C and N contents of the soil samples were determined using an elemental analyser (EA 2400 Series; Perkin Elmer, Foster City, CA, USA). For the determination of the inorganic-N concentrations (N[O.sup.-.sub.3] -N and N[H.sup.+.sub.4] - N), 3 g fresh soil, subsampled from each plastic bag, was extracted with 10% KCl (13 mL). After shaking (0.5 h), the extradant was filtered through filter paper (Grade 5C, <5 [micro]m; Advantec, Tokyo, Japan). Then, inorganic-N concentrations in soil samples were measured using a colorimetric method with a flow injection analyser (AQLA-700; Aqualab, Tokyo, Japan) as described previously (Hamamoto and Uchida 2015). Results are expressed as mg N [g.sup.-1] soil.
Microbial analyses of sampled soils
Soil DNA was extracted from each soil sample using the Powersoil DNA extraction kit (MoBio Laboratories, Carlsbad, CA, USA) according to the manufacturer's instructions. Fresh soil was used for DNA extraction and, for each soil sampling plot, three DNA samples were taken. The Ion Torrent system (Life Technologies, Rockville, MD, USA) was used for I6S rRNA analyses. The extracted DNA was amplified targeting the V2-4-8 and V3-6,7-9 regions of the 16S rRNA gene using the Ion 16S Metagenomics kit (Life Technologies 2014). The polymerase chain reaction (PCR) sample contained 15 [micro]L of 2x Environmental Master Mix (Ion Torrent; Life Technologies), 3 [micro]L of 16S primer (Ion Torrent; Life Technologies) and 1 [micro]L DNA extract and was made up to a final volume of 30 [micro]L using nuclease-free water. The PCR cycling conditions were 600 s at 95[degrees]C, followed by 25 cycles of 30 s at 95[degrees]C, 30 s at 58[degrees]C and 20 s at 72[degrees]C, with a final extension of 420 s at 72[degrees]C. The PCR products were quantified using the Qubit ds DNA HS Assay Kit (Invitrogen, Carlsbad, CA, USA). Then, 10-100 ng PCR products underwent a phosphorylation process with 1 [micro]L End Repair Enzyme (Ion Torrent; Life Technologies) and 20 [micro]L of 5x End Repair Buffer (Ion Torrent; Life Technologies) in total volume of 100 [micro]L, followed by a purification process using the Agencourt AMPure XP Kit (Beckman Coulter, Fullerton, CA, USA). The libraries were then attached with Ion PI adaptor (Ion Torrent; Life Technologies) and each barcode to make Ion Torrent sequencing sample specific. The PCR samples contained 12.5[micro]L DNA, 5[micro]L of 10x Ligase Buffer, 1 [micro]L Ion PI Adaptor, 1 [micro]L Ion Xpress Barcode, 1[micro]L dNTP mix, 1 [micro]L DNA ligase and 4 [micro]L, Nick Repair Polymerase, provided in the Ion Plus Fragment Library Kit and Ion Xpress Barcode Adapters 1-32 Kit (Life Technologies), and were made up to 50|iL using nuclease-free water. The PCR conditions were 900 s at 25[degrees]C and 300 s at 72[degrees]C. The libraries were purified using the Agencourt AM Pure XP Kit (Bcckman Coulter). Next, samples were re-amplified to increase the template successfully attached with the barcode. The volume of the PCR samples was 65 [micro]L, consisting of 50[micro]L Platinum PCR SuperMix High Fidelity (Ion Torrent; Life Technologies), 2.5 [micro]L Library Amplification Primer Mix (Ion Torrent; Life Technologies) and 10 [micro] L DNA sample filled with Low TE (Tris-EDTA; 10 mM Tris base, 0.1 mM EDTA, Ion Torrent; Life Technologies). The PCR cycling conditions were 300 s at 95[degrees]C, followed by six cycles of 15s at 95[degrees]C, 15s at 58[degrees]C and 60s at 70[degrees]C. The libraries were purified using the Agencourt AMPure XP Kit (Beckman Coulter). The final length and concentrations of the amplicons were checked using a Bioanalyzer High Sensitivity DNA Kit (Agilent Technologies, Palo Alto, CA, USA). Then, the library was diluted to 26 pM and set up on the OneTouch 2 instrument (Ion Torrent; Life Technologies) with the Ion PGM Template OT2 400 kit according to the manufacturer's instructions. The OneTouch ES system (Ion Torrent; Life Technologies) was used to purify the templated beads. Sequencing was performed using the Ion 316 chip with an Ion PGM 400 kit, and data were analysed online using Torrent Suite Software V5.0 (16S Metagenomics workflow V5.0, Thermo Fisher Scientific, Waltham, MA, USA).
Laboratory experimental set-up for the measurement of decreasing N[O.sup.-.sub.3]-N
Using the CF and HV soils sampled from the tomato greenhouse, an incubation experiment was performed to investigate differences in the rate of N[O.sup.-.sub.3]-N decrease with the addition of C and N. Fresh soils were placed in glass bottles (100 g dry soil per bottle) and were treated with one of the four N fertiliser and/or rice straw (C source) application treatments: (1) addition of straw and N; (2) addition of straw but not N; (3) no straw, but the addition of N; and (4) no straw, no N addition. The experimental set-up consisted of three replicates of two soil types (CF and HV), with or without N application (1 g N [kg.sup.-1] soil as KN[O.sub.3]-N), with or without straw application (100g rice straw [kg.sup.-1] soil; mean[+ or -]s.d. C:N ratio 94.5 [+ or -] 6.6). Water content was adjusted to 0.6 g water [g.sup.-1] dry soil at the commencement of the experiment. Water was added every day to maintain the water content during incubation. The original soil moisture content was determined to maximise microbial activity (Liang et al. 2003). Mean ([+ or -] s.d.) soil bulk density throughout the experiment for soils with and without straw was 0.80 [+ or -]0.10 and 0.48 [+ or -] 0.03 g [cm.sup.-3] respectively. Thus, the percentage of water-filled pore space (WFPS) for soils without and with ricc straw was 69% and 35% respectively. Incubations were conducted at room temperature (~25[degrees]C). From each bottle, soil subsamples (5 g) were taken on Days 1, 3, 5, 7, 9 and 14 after the start of incubation for the analyses described below.
To determine the decrease in N[O.sup.-.sub.3]-N, concentrations of N[O.sup.-.sub.3]-N and N[H.sup.+.sub.4]-N were measured as described above using the methods of Hamamoto and Uchida (2015). To determine changes in the microbial biomass C of the subsampled soils, a chloroform fumigation method was used. Briefly, two 1-g samples were taken from each subsampled soil; one was immediately extracted with 0.5 M [K.sub.2]S[O.sub.4] (10mL) and filtered through filter paper (Advantec Grade 5C), whereas the other 1-g sample was fumigated under chloroform for 48 h using an evacuated desiccator and then extracted in the same way. The extracts were stored in a freezer until C analysis. At that point, extracts were diluted between 1:2 and 1:10 with water and organic C was determined in 25-mL aliquots of the diluted samples using a total organic C analyser (TOC-5000A; Shimadzu Oceania, Sydney, NSW, Australia). The amount of organic C in the non-fumigated samples was subtracted from the amount of organic C in the fumigated samples; then, values were multiplied by 2.64 to convert them to microbial biomass C (Vance et al. 1987).
For subsamples taken on Days 1, 5 and 14 during the incubation experiment, soil bacterial community structures were determined using the 16S rRNA-based method as described above for the microbial analysis of sampled soil. For this sequencing measurement, one Ion 314 chip (Thermo Fisher Scientific) was used for 18 DNA samples (two soil types x three time points (Days 1, 5 and 14) x three replicates; see Table S2, available as Supplementary Material to this paper).
For the microbial analysis of sampled soil, the Shannon diversity index was calculated on the basis of 16S rRNA data using the Vegan package in R software version 3.1.2 (R Foundation for Statistical Computing, Vienna, Austria). The significance of differences between the two soils (in the Shannon diversity index) were evaluated using t-tests and were considered significant at two-tailed P< 0.05. To determine the significance of differences in 16S rRNA bacterial community structures between the two soils, a correspondence analysis was performed based on weighted Chi-squared distance matrices of their bacterial communities at the phylum level using the cca function in R software. Differences between the ratio of individual microbial phyla in the two soils were investigated using t-tests. For the incubation experiment, two-way analysis of variance (ANOVA) was used to determine the effects of soil history (long-term use of CF or HV application) and sampling date on soil N[O.sup.-.sub.3]-N concentration and biomass C. Unless indicated otherwise, data arc presented as the mean [+ or -] s.d.
Microbial analyses of sampled soils
On average, 66 631 reads were mapped per sample for 16S rRNA in the two soils (Table SI). Some 16S rRNA samples with a low mapped read number (< 10 000 reads) and samples that failed to amplify with the primers within the Ion 16S Metagenomics Kit were removed before further analysis. In all, eight samples for the CF soil and seven samples for the HV soil proceeded to further analysis. Correspondence analyses revealed significant differences between the two soil types in terms of 16S rRNA-based bacterial community composition (P<0.05; Fig. 1; Fig. S1). For both soil types, there were six predominant phyla: Proteobacteria, Actinobacteria, Acidobacteria, Cyanobacteria, Chloroflcxi and Gemmatimonadetes (Fig. 2). Actinobacteria and Cyanobacteria were more predominant in CF soil (46 [+ or -] 8% and 1.7 [+ or -] 0.4% respectively) than in HV soil (34 [+ or -]6% and 0.7 [+ or -]0.3% respectively), whereas Proteobacteria and Gemmatimonadetcs were more predominant in HV (47 [+ or -]7% and 2.2 [+ or -]0.9% respectively) than CF (37 [+ or -] 5% and 1.1 [+ or -] 0.6% respectively) soil (Fig. 2). The proportion of other phyla (Acidobacteria and Chloroflexi) was similar in the two soil types. The Phylum Proteobacteria was primarily composed of the Classes Alphaproteobacteria, Betaproteobacteria, Deltaproteobacteria and Gammaproteobacteria (23 [+ or -]4%, 8.7 [+ or -]2.8%, 2.6 [+ or -]0.9% and 2.9 [+ or -]1.1% respectively in CF soil; 18 [+ or -]3%, 14 [+ or -]6%, 4.9 [+ or -]1.4% and 8.9 [+ or -]4.2% respectively in HV soil; Fig. 2). Actinobacteria consisted primarily of the Class Actinobacteria (>70% within this phylum; data not shown) and the Order Actinomycetales (>99% within the Class Actinobacteria; data not shown). Within the Order Actinomycetales, the Families Nocardioidaceae and Streptomycetaccae were predominant (9.1 [+ or -] 1.8% and 5.9 [+ or -]2.2% respectively in CF soil; 7.1 [+ or -]2.9% and 2.9 [+ or -] 1.2% respectively in HV soil; Fig. 2). Based on Shannon diversity indices, soil bacterial communities were more diverse in HV than CF soil at the family level (P<0.05), but not at the phylum, class or order level (Table 2).
After the addition of N[O.sup.-.sub.3]-N and straw, concentrations of N[O.sup.-.sub.3]-N in the CF and HV soils decreased rapidly (Fig. 3a), but these decreases were not seen in samples incubated without straw (Fig. 3c). Similar observations were made for soils incubated without N[O.sup.-.sub.3]-N, but the concentration of N[O.sup.-.sub.3]-N in these samples at the beginning of the experiment was much lower than N[O.sup.-.sub.3]-N concentrations in the soils with added N[O.sup.-.sub.3]-N (Fig. 3b, d). The soils without N[O.sup.-.sub.3]-N and without straw showed significant differences between the CF and HV soils in the N[O.sup.-.sub.3]-N concentrations (Fig. 3d). For the two soils with added C and N (Fig. 3a), there was an interaction between soil type and sampling period (day) and, in both soils, a decrease in N[O.sup.-.sub.3]-N within 7 days was observed. The concentration of N[H.sup.+.sub.4]-N in the two soils was similar during the incubation, and the amount of N[H.sup.+.sub.4]-N was much smaller than that of N[O.sup.-.sub.3]-N (Figs 3, 4).
For soil microbial biomass C, there were no clear differences between the soils except for the control plots, in which a relatively higher soil microbial biomass C was found in HV than CF soils (Fig. 5). Soil microbial biomass C tended to be larger in soils with an additional C source (Fig. 5a, b) compared with those in which a C source was not added (Fig. 5c, d). For the soils with C and N addition (Fig. 5a), microbial biomass C peaked on Day 1 and decreased towards Day 9. However, at the final sampling (Day 14), microbial biomass C increased at levels above those on Day 9 but remained lower than on Day 1.
16S rRNA bacterial community structures were investigated in soils with C and N addition, with marked changes in rRNA community structures observed during the incubation period (Fig. S2). In addition, the proportion of the four predominant phyla (Proteobacteria, Firmicutes, Actinobacteria and Bacteroidetes) within the bacterial communities changcd during the period of incubation with C and N (Fig. 6). On Day 1 after the addition of C and N, the CF soils had a relatively larger proportion of Actinobacteria and Firmicutes and smaller proportion of Bacteroidetes and Proteobacteria than the HV soils. On Day 5, the proportion of Proteobacteria became larger in CF soils than in HV soils (P<0.05). On Day 14, only Actinobacteria differed significantly between the CF or HV soils (P<0.05). The Shannon diversity index suggested that, in general, soil bacterial communities were more diverse in HV soils (Table 3). However, at the end of the experiment (Day 14, Table 3), the difference in the Shannon diversity index between the CF and HV soils was no longer significant (Table 3). In the incubation experiments, the difference between the diversities of CF and HV decreased over time after the addition of C and N.
Soil microbial structures after the long-term use of CF and HV
In the present study we observed greater soil microbial biomass C in the HV than CF soil during incubation of the soils without any addition of C and N (Fig. 5). This suggests that the microbial biomass increased in soils with a history of HV rather than CF application. In addition, soil bacterial community structures differed significantly between the two soils with different management histories (Fig. 1). Actinobacteria was the predominant phylum in CF soil (Fig. 2). This phylum is composed of several classes, but in both soils the most common organisms were in the Class Actinobacteria, Order Actinomycetales and Families Nocardioidaceae and Streptomycetaceae. Within these two families, the proportion of Streptomycetaceae was higher in CF than HV soil (Fig. 2), thus the long-term application of CF may have increased this family or suppressed other similar families within the phylum. Bonilla et al. (2016) showed that the growth of Streptomycetaceae was increased in culture medium following the addition of 5 mM sulfate. Fischer et al. (2012) also showed that the Genus Streptomyces in Streptomycetaceae uses sulfate as a sulfur (S) source. In the present case, ammonium sulfate was applied as the fertiliser for the CF soil every year. Thus, the increased S in the CF compared with H V soil may have been one of the reasons for the increase in Actinobacteria in the CF soil seen in the present study. Another significant phylum in the CF soil was Cyanobacteria, which is a nitrifier and contains nitric oxide reductase within its genome (Busch et al. 2002). The amount of total N was significantly greater in CF soil (Table 1). Thus, Cyanobacteria involved in nitrification may have been stimulated in CF soil.
The proportion of Gemmatimonadetes and Proteobacteria was significantly higher in the HV than CF soil (Fig. 2). Bernard et al. (2007) showed that Gemmatimonadetes was involved in the decomposition of recalcitrant soil organic matter rather than fresh organic matter, indicating that this phylum has the ability to use a more recalcitrant C source. In addition, based on stable isotope probing, Lee et al. (2011) showed that Gemmatimonadetes assimilated C from rice callus. Thus, Gemmatimonadetes may respond positively to a recalcitrant C source derived from HV. Another phylum predominant in our soils was Proteobacteria, and this phylum is composed of several classes that have generally been described as copiotrophic taxa favoured by nutrient-rich conditions (Bernard et al. 2007; Fierer et al. 2007; Lee et al. 2011). Fierer et al. (2007) showed that the abundance of Betaproteobacteria was positively correlated with the C mineralisation rate and Baneijee et al. (2016) showed that the abundance of Gammaproteobacteria increased more than 10-fold in soils after the addition of wheaten straw and the nutrients N, P and S. Furthermore, Ridl et al. (2016) reported that the presence of plants enhanced the abundance of Beta-, Gamma- and Deltaproteobacteria compared with non-vegetated soils. Similarly, in the present study, the proportion of Protcobacteria was significantly higher in HV than CF soil, indicating a higher proportion of Beta-, Gamma- and Deltaproteobacteria in the HV soil (Fig. 2).
Based on values of the Shannon diversity index, the present study suggests that the long-term use of HV did not change bacterial diversity at the phylum level, but did increase it at the family level (Table 2). Some previous studies have reported no change in bacterial diversity at the family level following crop residue return to the soil (Coudrain et al. 2016) or at the phylum level in response to organic farming management (Orr et al. 2015) based on 16S rRNA analysis. Further investigations are needed in this area.
Changes in soil nitrate concentrations and microbial biomass C after C and N addition
Despite the differences in bacterial community structure between the two soils, the rate of N03 decrease after the addition of C and N was similar, with [N.sub.3] decreasing rapidly over a period of 10 days. Furthermore, there was no change in the N[H.sup.+.sub.4]-N concentration throughout the incubation period, with the N[H.sup.+.sub.4]-N concentration being very low (Fig. 4). Thus, the nitrification process apparently did not affect the rate of N[O.sup.-.sub.3] decrease in the present study. We did not measure denitrification, but because of the rather small (35%) WFPS in the soils with rice straw, we believe that denitrification was not the major reason for the rapid decrease in N[O.sup.-.sub.3] in the present study (Bateman and Baggs 2005; Jetten 2008; Saggar et al. 2013). Further studies should focus more on functional genes. such as nitrate reductase (nirK, nirS) and nitrous oxide reductase (nosZ), which are responsible for the denitrification processes (Wang et al. 2017). Considering the increase in biomass C (Fig. 5a), it is likely that N[O.sup.-.sub.3] was used by soil microbes after the addition of both C and N. The amount of N03 did not change when only excess N was added (Fig. 3c), suggesting that N assimilation by microbes in these soils was limited by C.
We did not measure the amount of C decomposed, thus we do not know whether the two soils had different capacities to decompose C. However, previous studies showed that C addition enhanced C mineralisation, followed by an increase in microbial biomass C (Bingeman et al. 1953; Fontaine et al. 2004; Qiu et al. 2016). Fontaine et al. (2004) also reported an increase in microbial biomass C resulting from the assimilation of added [sup.13]C-cellulose; however, within 13 days, 85% of the labelled C was decomposed. Thus, in the present study, the increase in microbial biomass C immediately after the addition of excess C and N (Fig. 5a) may be due to straw C, but the temporal increase was followed by a decrease within 9 days, apparently due to decreased availability of N and readily decomposable C. Moreno-Cornejo et al. (2015) reported an increase in microbial biomass C immediately after the addition of plant residues, followed by a rapid decrease. Qiu et al. (2016) also reported that soil organic matter decomposition was stimulated for 7 days after maize residue addition. Interestingly, Moreno-Cornejo et al. (2015) found an increase in microbial biomass C at the later stages of plant residue decomposition, suggesting incorporation of recalcitrant C by slow- growing microbes. Therefore, the results of the present study are in agreement with those of previous studies, and the increase in microbial biomass C between Days 9 and 14 after the addition of C and N (Fig. 5a) may suggest the decomposition of more recalcitrant C from straw.
Bacterial community responses to C and N input
When C and N were added to the CF and HV soil, the responses of Bacteroidetes, Firmicutes and Proteobacteria were significantly different on day I (Fig. 6a). However, these differences disappeared as the incubation progressed (Fig. 6b and c). In addition, the difference in the Shannon diversity index between the CF and HV soils was no longer significant after 14 days (Table 3). These results suggest that even though there were significant differences in soil microbial community structure and diversity at the beginning of the incubation period, the addition of excess C and N quickly reduced differences in soil microbial community structure.
During the 14-day incubation period with excess C and N, there were increases in Proteobacteria and Bacteroidetes over time, whereas Firmicutes was undetected except on Day I (Fig. 6). Similar to Firmicutes, levels of Actinobacteria also decreased over time. The microbial changes in this experiment confirm the results of previous studies. Pascault et al. (2013) showed that Firmicutes was predominant 3 days after the addition of organic matter, but was replaced by Proteobacteria after
14 days. Furthermore, at 60 days, Proteobacteria was the main phylum comprising the residue-degrading community (Pascault et al. 2013). In another study, Banetjee et al. (2016) reported that the addition of nutrients (straw and N, P and S) increased the abundance of Gammaproteobacteria but reduced the abundance of Firmicutes over a period of 14 days. Proteobacteria, Bacteroidetes and Firmicutes, which became predominant after C and N addition, arc described as taxa favoured by nutrient availability (Fierer et al. 2007; Van Horn et al. 2014; Leff et al. 2015).
The changes in microbial composition over 14 days can be explained by different growth rates for each bacterial phylum. Growth rates differ among bacterial phyla, with some being described as fast growing and others as slow growing (Bernard et al. 2007; Pascault et al. 2013). Thus, the increase in Firmicutes in the early stages of incubation and the increases in Proteobacteria and Bacteroidetes at later stages may indicate a faster growth of Firmicutes than of Proteobacteria and Bacteroidetes. The ability of each phylum to use C also differs. Pascault et al. (2013) showed that Proteobacteria was predominant in soil amended with alfalfa residue, an easily decomposed C source, whereas Firmicutes predominated in soils amended with wheat residues, a recalcitrant C source. Actinobacteria is adaptable to many kinds of C sources (Lewin et al. 2016). Thus, in our experiment, Firmicutes and Actinobacteria predominated in the early stages of incubation because they could use the readily available C source from rice straw. In contrast, Proteobacteria and Bacteroidetes used the more recalcitrant C remaining after C decomposition by Firmicutes and Actinobacteria.
The use of HV for 9 years as a green manure before growing tomatoes in a greenhouse soil significantly changed the microbial community structure of the soil based on 16S rRNA analysis compared with soils conventionally managed with chemical N fertiliser. Actinobacteria and Cyanobacteria were predominated in the CF-treated soil, whereas Proteobacteria and Gemmatimonadetes predominated in the HV-treated soil. The structural diversity of the soil bacterial community was also changed at the family level by the use of HV compared with CF, but this was not observed at the phylum level. In addition, differences in soil bacterial community structures did not affect the rate of soil nitrate decrease when excess C (rice straw) and N (nitrate) were applied to the soils. The addition of C and N significantly decreased the diversity of soil microbes within 14 days and the differences in the community structures between the two soils were no longer significant. This may be due to the marked increase in Proteobacteria and Bacteroidetes and the rapid increase followed by decline in Firmicutes. Considering these points, even though community structure and diversity at the family level were altered by the long-term use of HV as green manure, soil function, as measured by the rate of decrease in nitrate concentrations with rice straw added as a C source, was not altered over the short term.
Information on the read number during the DNA analyses and information of the bacterial phylum for each replicate arc available from the Journal's website.
This research was fiinded by grants from MEXT/JSPS KAKENHI (JP16K18663 and JP26520301). The authors thank Dr Chiaki Hon and Ms Reika Isoda for their assistance with the microbial experiments.
Araki H, Hane S, Hoshino Y, Hirata T (2009) Cover crop use in tomato production in plastic high tunnel. Horticulture, Environment and Biotechnology 50, 324-328.
Banerjee S, Kirkby CA, Schmutter D, Bissett A, Kirkegaard JA, Richardson AE (2016) Network analysis reveals functional redundancy and keystone taxa amongst bacterial and fungal communities during organic matter decomposition in an arable soil. Soil Biology & Biochemistry 97, 188-198. doi: 10.10I6/j.soilbio.2016.03.017
Barrett JE, Burke IC (2000) Potential nitrogen immobilisation in grassland soils across a soil organic matter gradient. Soil Biology & Biochemistry 32, 1707-1716. doi: 10.1016/S0038-0717(00)00089-4
Bateman EJ, Baggs EM (2005) Contributions of nitrification and denitrification to [N.sub.2]O emissions from soils at different water-filled pore space. Biology and Fertility of Soils 41, 379-388. doi: 10.1007/ S00374-005-0858-3
Bernard L, Mougel C, Maron P, Nowak V, Leveque J, Henault C, Haichar FZ, Berge O, Marol C, Balesdent J, Gibiat F, Lemanceau P, Ranjard L (2007) Dynamics and identification of soil microbial populations actively assimilating carbon from 13C-labelled wheat residue as estimated by DNA- and RNA-S1P techniques. Environmental Microbiology 9, 752-764. doi:10.111 l/j.1462-2920.2006.01197.x
Bingeman CW, Varner JE, Martin WP (1953) The effect of the addition of organic materials on the decomposition of an organic soil. Soil Science Society of America Journal 29, 692-696.
Blasko R, Hogberg P, Bach LH, Hogberg MN (2013) Relations among soil microbial community composition, nitrogen turnover, and tree growth in N-loaded and previously N-loaded boreal spruce forest. Forest Ecology and Management 302, 319-328. doi:10.1016/j.foreco. 2013.02.035
Bonilla JO, Callegari EA, Delfini CD, Estevez MC, Villegas LB (2016) Simultaneous Chromate and sulfate removal by Streptomyces sp. MCI. Changes in intracellular protein profile induced by Cr(VI). Journal of Basic Microbiology 56. 1212 1221.
Busch A, Friedrich B, Cramm R (2002) Characterisation of the norB gene, encoding nitric oxide reductase, in the nondenitrifying cyanobacterium Synechocystis sp. strain PCC6803. Applied and Environmental Microbiology 68, 668-672. doi: 10.1128/AEM.68.2.668-672.2002
Chen B, Liu E, Tian Q, Yan C, Zhang Y (2014) Soil nitrogen dynamics and crop residues. A review. Agronomy for Sustainable Development 34, 429-442. doi: 10.1007/sl 3593-014-0207-8
Coudrain V, Hedde M, Chauvat M, Maron P, Bourgeois E, Mary B, Leonard J, Ekelund F, Villenave C, Recous S (2016) Temporal differentiation of soil communities in response to arable crop management strategies. Agriculture, Ecosystems and Environment 225, 12-21.
de Vries FT, Bardgett RD (2012) Plant-microbial linkages and ecosystem nitrogen retention: lessons for sustainable agriculture. Frontiers in Ecology and the Environment 10, 425-432. dot:10.1890/110162
Fierer N, Bradford MA, Jackson RB (2007) Toward an ecological classification of soil bacteria. Ecology 88. 1354 1364. doi: 10.1890/ 05-1839
Fischer M, Schmidt C, Falke D, Sawers RG (2012) Terminal reduction reactions of nitrate and sulfate assimilation in Streptomyces coelicolor A3(2): identification of genes encoding nitrite and sulfite reductases. Research in Microbiology 163, 340-348. doi:10.1016/j.resmic.2012. 05.004
Fontaine S, Bardoux G, Benest D, Verdier B, Mariotti A, Abbadie L (2004) Mechanisms of the priming effect in a savannah soil amended with cellulose. Soil Science Society of America Journal 68, 125--131. doi : 10.2136/sssaj2004.1250
Hamamoto T, Uchida Y (2015) Sodium contents in dairy cow urine and soil aggregate sizes influence the amount of nitrogen lost from soil. Applied and Environmental Soil Science 2015, 275985. doi: 10.1155/2015/275 985
Jetten MSM (2008) The microbial nitrogen cycle. Environmental Microbiology 10, 2903-2909. doi:10.1111/j.1462-2920.2008.01786.x
Lee CG, Watanabe T, Sato Y, Murase J, Asakawa S, Kimura M (2011) Bacterial populations assimilating carbon from [sup.13]C-labeled plant residue in soil: analysis by a DNA-SIP approach. Soil Biology & Biochemistry 43. 814-822. doi: 10.1016/j.soilbio.2010.12.016
Leff JW, Jones SE, Prober SM, Barberan A, Borer ET, Firn JL, Harpole WS, Hobbie SE, Hoftnockel KS, Knops JMH, McCulley RL, La Pierre K, Risch AC, Seabloom EW, Schutz M, Steenbock C, Stevens CJ, Fierer N (2015) Consistent responses of soil microbial communities to elevated nutrient inputs in grasslands across the globe. Proceedings of the National Academy of Sciences of the United States of America 112, 10967-10972. doi : 10.1073/pnas.1508382112
Lewin GR, Carlos C, Chevrette MG, Horn HA, McDonald BR, Stankey RJ, Fox BG, Currie CR (2016) Evolution and ecology of Actinobacteria and their bioenergy. Applied Microbiology 70, 235-254.
Liang C, Das KC, McClendon RW (2003) The influence of temperature and moisture contents regimes on the aerobic microbial activity of a biosolids composting blend. Bioresource Technology 86, 131 137. doi: 10.1016/S0960-8524(02)00153-0
Life Technologies (2014) Ion Torrent. Application note. 16S rRNA sequencing. An integrated research solution for bacterial identification using 16S rRNA sequencing on the Ion PGM[TM] System with Ion Reporter[TM] Software. Available at https://www.thermofisher.com/content/dam/LifeTech/Docu ments/PDFs/Ion-16S-Metagenomics-Kit-Software-Application-Note.pdf [verified 17 April 2017],
Mao Y, Yannarell AC, Mackie RI (2011) Changes in N-transforming archaea and bacteria in soil during the establishment of bioenergy crops PLOS One 6, e24750. doi:10.1371/journal.pone.0024750
Mary B, Recous S, Darwis D, Robin D (1996) Interactions between decomposition of plant residues and nitrogen cycling in soil. Plant and Soil 181, 71-82. doi: 10.1007/BF00011294
Moreno-Comejo J, Zornoza R, Doane TA, Faz A, Horwath WR (2015) Influence of cropping system management and crop residue addition on soil carbon turnover through the microbial biomass. Biology and Fertility of Soils 51, 839-845. doi:10.1007/s00374-015-1030-3
Orr CH, Stewart CJ, Leifert C, Cooper JM, Cummings SP (2015) The effect of the addition of organic materials on the decomposition of an organic soil. Journal of Applied Microbiology 119. 208-214. doi: 10.1111/jam. 12822
Orwin KH, Buckland SM, Johnson D, Turner BL, Smart S, Oakley S, Bardgett RD (2010) Linkages of plant traits to soil properties and the functioning of temperate grassland. Journal of Ecology 98, 1074-1083. doi: 10.1111/j.1365-2745.2010.01679.x
Pascault N, Ranjard L, Kaisermann A, Bachar D, Christen R, Terrat S, Mathieu O, Leveque J, Mougel C, Henault C, Lemanceau P, Pe'an M, Boiry S, Fontaine S, Maron P (2013) Stimulation of different functional groups of bacteria by various plant residues as a driver of soil priming effect. Ecosystems 16, 810-822. doi: 10.1007/s10021-0139650-7
Qiu Q, Wu L, Ouyanga Z, Lia B, Xu Y, Wu S, Gregorich EG (2016) Priming effect of maize residue and urea N on soil organic matter changes with time. Applied Soil Ecology 100, 65-74. doi: 10.1016/j.apsoil.2015.11. 016
Ridl J, Kolar M, Strejcek M, Strnad H, Stursa P, Paces J, Macek T, Uhlik O (2016) Plants rather than mineral fertilisation shape microbial community structure and functional potential in legacy contaminated soil. Frontiers in Microbiology 7, 995. doi:10.3389/fmicb.2016. 00995
Saggar S, Jha N, Deslippe J, Bolan NS, Luo J, Giltrap DL, Kim D-G, Zaman M, Tillman RW (2013) Denitrification and [N.sub.2]O:N2 production in temperate grasslands: processes, measurements, modelling and mitigating negative impacts. The Science of the Total Environment 465, 173-195. doi: 10.1016/j.scitotenv.2012.11.050
Van Horn DJ, Okie JG, Buelow HN, Gooseff MN, Barrett JE, TakacsVesbach CD (2014) Soil microbial responses to increased moisture and organic resources along a salinity gradient in a polar desert. Applied and Environmental Microbiology 80, 3034-3043. doi: 10.1128/ AEM.03414-13
Vance ED, Brooks PC, Jenkinson DS (1987) An extraction method for measuring soil microbial biomass C. Soil Biology & Biochemistry 19, 703-707. doi: 10.1016/0038-0717(87)90052-6
Wang Y, Uchida Y, Shimomura Y, Akiyama H, Hayatsu M (2017) Responses of denitrifying bacterial communities to short-term waterlogging of soils. Scientific Reports 7, 803. doi: 10.1038/s41598-017-00953-8
Misato Toda (A) and Yoshitaka Uchida (B,C)
(A) Graduate School of Agriculture, Hokkaido University, Kita 9, Nishi 9, Kita-ku, Sapporo, Hokkaido 060-8589, Japan.
(B) Research Faculty of Agriculture, Hokkaido University, Kita 9, Nishi 9, Kita-ku, Sapporo, Hokkaido 060-8589, Japan.
(C) Corresponding author. Email: firstname.lastname@example.org
Received 17 April 2017, accepted 2 June 2017, published online 13 July 2017
Caption: Fig. 1. Multidimensional projection of soil samples using analyses based on weighted Chi-squared distance matrices of their bacterial communities. In the analyses, differences between different soil microbial community structures were expressed in the two-dimensional scatter plot. Thus, the CA1 (x-axis) and CA2 (y-axis) data do not have units but the distance between each point corresponds to the similarity of the two soil microbial community structures. The labels correspond to the soil types i.e. chemical fertiliser (CF) and hairy vetch (HV) soils. The difference between two soils was significant (P<0.05). Each plot represents one microbial DNA sample. The locations of the labels (CF and HV) are the centroids of replicated microbial community data.
Caption: Fig. 2. Relative abundance of dominant bacterial phyla showing significant effects (P<0.05) of soil history, namely a history of chemical fertiliser application (CF) or a history of hairy vetch application (HV) for 9 years. HV application had a positive effect on Gemmatimonadetes and Proteobacteria and a negative effect on Actinobacteria and Cyanobacteria. Data are the mean [+ or -] s.d. * P<0.05, **P<0.01 respectively.
Caption: Fig. 3. Changes in N[O.sup.-.sub.3]-N concentrations in soils with a history of chemical fertiliser (CF) or hairy vetch (HV) application (a) with carbon (C) and nitrogen (N) input, (ft) with C input, (c) with N input and (d) without C and N input (control) during the 14 days incubation.
Caption: Fig. 4. Change in N[H.sup.-.sub.4]-N concentrations in soils with a history of chemical fertiliser (CF) or hairy vetch (HV) application (a) with carbon (C) and nitrogen (N) input, (h) with C input, (c) with N input and (d) without C and N input (control) during the 14 days incubation.
Caption: Fig. 5. Change in microbial biomass carbon of soils with a history of chemical fertiliser (CF) or hairy vetch (HV) application (a) with carbon (C) and nitrogen (N) input, (b) with C input, (c) with N input and (d) without C and N input (control) during the 14 days incubation.
Caption: Fig. 6. Relative abundance of dominant bacterial phyla with C and N input over 14 days incubation: (a) 1 day; (b) 5 days; (c) 14 days. CF, soils with a history of chemical fertiliser application for 9 years; HV, soils with a history of hairy vetch application for 9 years. The number of phyla that is significantly different between two soil decreased over time during the incubation. Data are the mean [+ or -] s.d. *P<0.05.
Table 1. Properties for soils sampled in hairy vetch (HV) and chemical fertiliser (CF) plots from a tomato trial (n = 3) Data are the mean [+ or -] s.d. WFPS, water-filled pore space; NS, not significant (P-value>0.05) HV CF Water content (%) 17.0 [+ or -] 0.5 14 [+ or -] 2 pH ([H.sub.2]O) 6.9 [+ or -] 0.2 6.4 [+ or -] 0.8 WFPS (%) 0.27 [+ or -] 0.04 0.23 [+ or -] 0.04 Total C (g [kg.sup.-1]) 38.6 [+ or -] 0.2 39 [+ or -] 3 Total N (g [kg.sup.-1]) 3.4 [+ or -] 0.3 3.6[+ or -] 0.3 C : N ratio 12 [+ or -] 1 11 [+ or -] 1 N[O.sub.3.sup.-] N (mg [kg.sup.-1]) 201 [+ or -] 167 279 [+ or -] 45 N[H.sub.4.sup.+] N (mg [kg.sup.-1]) 4 [+ or -] 2 34 [+ or -] 47 p-value Water content (%) 0.02 pH ([H.sub.2]O) NS WFPS (%) NS Total C (g [kg.sup.-1]) NS Total N (g [kg.sup.-1]) 0.03 C : N ratio NS N[O.sub.3.sup.-] N (mg [kg.sup.-1]) NS N[H.sub.4.sup.+] N (mg [kg.sup.-1]) NS Table 2. Shannon's diversity index for chemical fertiliser (CF) and hairy vetch (HV) soils Data are the mean [+ or -] s.d.; NS, not significant (P-value>0.05) Soil type Phylum Class Order CF 1.3 [+ or -] 0.1 2.2 [+ or -] 0.2 2.6 [+ or -] 0.3 HV 1.4 [+ or -] 0.1 2.4 [+ or -] 0.0 3.0 [+ or -] 0.1 P-value NS NS NS Soil type Family n CF 3.6 [+ or -] 0.2 3 HV 4.0 [+ or -] 0.1 3 P-value <0.05 Table 3. Shannon's diversity index at the family level for chemical fertiliser (CF) and hairy vetch (HV) soils with carbon (C) and nitrogen (N) input over the 14-days incubation period Data are the mean [+ or -] s.d. NS, not significant (P-value > 0.05) Soil type Day 1 Day 5 Day 14 CF 1.6 [+ or -] 0.2 1.3 [+ or -] 1.1 2.1 [+ or -] 0.4 HV 2.2 [+ or -] 0.3 2.3 [+ or -] 0.2 1.9 [+ or -] 0.1 P-value <0.05 NS NS
|Printer friendly Cite/link Email Feedback|
|Author:||Toda, Misato; Uchida, Yoshitaka|
|Date:||Aug 1, 2017|
|Previous Article:||Can nitrogen fertiliser maintain wheat (Triticum aestivum) grain protein concentration in an elevated C[O.sub.2] environment?|
|Next Article:||Benchmarking and mitigation of nitrous oxide emissions from manures and fertilisers used in temperate vegetable crops in Australia.|